Regulation of Ovarian Follicular Development in Primates: Facts and Hypotheses

Transcription

1 X/96/$03.00/0 Endocrine Reviews Copyright 1996 by The Endocrine Society Vol. 17, No. 2 Printed in U.S.A. Regulation of Ovarian Follicular Development in Primates: Facts and Hypotheses ALAIN GOUGEON Institut National de la Sante et de la Recherche Medicale and Institut Paris-Sud sur les Cytokines, Clamart, France I. Introduction II. The Stock of Resting Follicles A. Depletion of the stock by follicular atresia B. Depletion of the stock by entry of follicles in growth phase: initiation of follicular growth 1. Morphometric studies 2. Factors triggering the initiation of follicular growth III. Morphological Aspects of Follicular Development A. Description of the follicular development in primates B. Cytological and morphometric changes of follicular tissues during development C. Dynamics of human follicular growth IV. Gonadotropins and Follicular Functions A. The primary role of gonadotropins to initiate and sustain follicular functions 1. Gonadotropins sustain follicular growth 2. Gonadotropins sustain follicular steroidogenesis B. Changes in follicular responsiveness to gonadotropins during folliculogenesis 1. The basal follicular growth 2. The selectable follicles 3. The newly selected follicle 4. Maturation of the preovulatory follicle V. Factors Regulating the Progressive Acquisition of Follicular Responsiveness to Gonadotropins A. The gonadotropin receptors B. The intraovarian peptides 1. The insulin-like growth factor (IGF) system 2. Epidermal growth factor (EGF) and transforming growth factor-a (TGFa) 3. Basic fibroblast growth factor (bfgf) 4. TGFJ3 5. Inhibin, activin, and follistatin (FSP) 6. Cytokines VI. Atresia A. Morphological characterization of atresia B. Follicular atresia in the human ovary during the menstrual cycle C. Alterations of follicular metabolism during atresia D. Causes of atresia Address reprint requests to: Alain Gougeon, INSERM U-355, 32 rue des Carnets, Clamart, France. 1. Mechanisms leading to activation of the apoptotic cascade 2. Alteration of the tropic hormonal signal 3. Other possible causes of atresia VII. Synthetic Views on the Control of Follicular Development in Primates A. Entry of follicles into the preantral class during the early luteal phase: physiological implications 1. Which factors are involved in the increased vascularization of preantral follicles? 2. Is there a coincidence between the periovulatory period and the differentiation of TICs? 3. Do changes represent protective mechanisms of a particular wave of follicular growth? B. Basal follicular growth 1. How do follicles develop during basal growth despite their low responsiveness to FSH? 2. Why is steroidogenic activity low in follicles during basal growth? C. Selectable follicles 1. How do these follicles respond to changing levels of FSH by altering GC proliferation? 2. Why does production of E 2 remain low in selectable follicles? D. Selection of the follicle destined for ovulation 1. What regulates changes in steroidogenic activities in the newly selected follicle? 2. How is the follicle destined to ovulate "chosen"? Are these follicles different because they are derived from a different clonal pool? E. Preovulatory maturation 1. What is the role of the inhibin-activin-fsp system during preovulatory maturation? 2. Why does aromatase activity increase before the midcycle gonadotropin surge? 3. Why does GC proliferation decrease before the gonadotropin surge? Is this change due to GC differentiation? 4. How does the preovulatory follicle exert its dominance upon less developed follicles? 5. How does the preovulatory follicle continue to thrive when FSH levels are decreasing? F. Preovulatory growth after the midcycle gonadotropin surge 1. What controls the shift in preovulatory steroidogenesis after the gonadotropin surge? 121

2 122 GOUGEON Vol. 17, No Which mechanism arrests mitoses in GCs after the gonadotropin surge? VIII. Conclusions and Perspectives I. Introduction IN THE adult ovary, folliculogenesis starts when follicles leave the pool of resting follicles (RF) to enter the growth phase. From there, the early growing follicle undergoes a developmental process including a dramatic course of cellular proliferation and differentiation. In primates, only one follicle commonly reaches the preovulatory stage every cycle; most follicles fail to complete this maturation scheme, dying in the process termed atresia. In recent years, a picture has emerged depicting the classic endocrine control of ovarian function by LH and FSH, entangled in a maze of regulatory systems hinging on cell-cell interactions between follicular cells, via action of a variety of molecules (1-3). Different types of cell-cell interactions have been described. In paracrine regulations, a molecule synthesized by one cellular type is released into the interstitial milieu to act directly on another cellular type. In autocrine regulations, molecules synthesized by one cellular type are released to act either on the same tissue (autocrine action), on immediate neighbor cells (juxtacrine action), or are not released and act inside the producing cell itself (intracrine action). These paracrine and autocrine regulations may be involved at every stage of folliculogenesis by acting either alone or by modifying the multiple functions of FSH and LH. Finally, it appears from these findings that during folliculogenesis a finely tuned process of progressive differentiation occurs in all constituents of the follicle. Recent articles have described the anatomic, physiological, and endocrine changes that occur within the ovary during the female reproductive cycle (4-7). In addition, many in vitro experiments describing actions of various types of molecules on follicular cell function have provided new clues as to the regulations involved in folliculogenesis. Despite the new and pertinent insights recently obtained, data are sometimes confusing or contradictory, and the in vitro studies are limited. These limitations include: 1) sera used to improve culture conditions contain various factors that interact with the studied tissue, possibly leading to paradoxical or artifacrual results; 2) the precise cell to cell interactions related to the structural organization of the follicle cannot be maintained; 3) any functional differences of individual cells within a cell type are obscured; and 4) because they are too small to be subjected to dissection and cell preparation procedures, resting, secondary, and preantral follicles are not easily investigated, especially in women with very low density due to both the size of the organ and the age of the subject. In addition, transposition of these data to primates in vivo must be performed with extreme caution since many of the findings are from various animal models with folliculogenesis patterns that are different from primates. For these reasons, the mechanisms that control in vivo ovarian folliculogenesis in primates remain poorly understood. Since it is now possible to quantify certain biological activities and localize enzymes, hormones, and intraovarian peptides as well as their receptors in/on follicular cells, one of the approaches taken to acquire additional insight into the actual behavior of follicles is to evaluate in situ their functional status at various times during the ovarian cycle. Consequently, in an attempt to build a dynamic model of folliculogenesis, these functional data have been analyzed and integrated in the light of the well known cyclic endocrine changes and of more hypothetical intraovarian / extraovarian regulations occurring at each given stage of follicular development. Thus, the aim of this review is to use a chronological perspective to review follicular development in primates, beginning with the initiation of follicular growth and ending with preovulatory changes occurring in the dominant follicle. The putative control mechanisms that may operate in vivo to ensure the successful maturation of a single preovulatory follicle have been particularly emphasized. Most of the concepts presented derive from studies in humans and subhuman primates. When appropriate data were not available in primates, extrapolations were made from data obtained in rodents and farm animals. II. The Stock of Resting Follicles In the human ovary, follicles begin to form during the fourth month of fetal life (8). Although some of these newly formed follicles start to grow almost immediately, most of them remain in a resting stage until they either degenerate or some signal(s) activate(s) them to enter the growth phase. These follicles constitute the stock of RF. Three types of RF (primordial, intermediary, small primary) have been defined in humans (Fig. 1). The RF stock is mainly comprised of primordial and intermediary follicles, which are seen in similar numbers in the right and left ovaries in both humans (9) and monkeys (10). At birth, the human ovary contains from 266,000 to 472,000 RF (11-13). With increasing age, the population of RF progressively decreases, leading to a stock at menopause estimated between less than 100 (13) and 1000 (14) RF. The age-related depletion of the RF stock occurs as a result of two processes: atresia and entry in growth phase. In the human ovary, the percentage of atretic RF, which has been estimated to be about 50% at birth (8), decreases from birth to approximately 30 yr of age (13). Up to 30 yr of age, the loss of RF is mainly due to RF atresia; thereafter, this loss of RF is due mainly to entrance of RF in growth phase. The sequence whereby the RF pool first decreases by atresia and then by entrance into the growth phase has been observed also in rodents during puberty and adult life, respectively (15, 16). The decay rate of RF accelerates from approximately 38 yr of age onward and is either due to an accelerated entry of RF in growth phase (13) or to an increasing atresia of RF (17). The discrepancy between these two studies can be explained by the follicular classification used by the authors, in which intermediary and small primary follicles were considered either as RF (13) or growing follicles (17). A. Depletion of the stock by follicular atresia Eosinophilia of the ooplasm, contraction followed by clumping of the chromatin material, and wrinkling of the oocyte nuclear membrane have been considered signs of RF atresia in human fetal and neonatal ovaries (8).

3 April, 1996 OVARIAN FOLLICULAR DEVELOPMENT IN PRIMATES 123 FIG. 1. Resting follicles in the adult human ovary, X570. A, Primordial follicle in which the oocyte is surrounded by flattened GCs. B, Intermediary follicle in which the oocyte is surrounded by a mixture of flattened and cuboidal GCs. C, Primary follicle in which the oocyte is surrounded by a single layer of cuboidal GCs. [Reproduced with permission from A. Gougeon, The Ovary, pp (5).] There are no data concerning the factors subserving atresia of RF in primates. In postnatal normal rodents, plasma FSH is high at birth and between 7 and 18 days of life (18), and during the first 2 weeks of life the RF stock is halved (19). Using rodents showing natural (20, 21) or experimentally (22) low circulating levels of gonadotropins, it has been suggested that in prepubertal animals postnatal gonadotropins prompted follicular loss by atresia. Yet, nonphysiological levels of gonadotropins (23, 24), hemicastration (25), and ovarian transplantation in a gonadotropin-rich environment (26) have all failed to accelerate RF atresia. Alternatively, postnatal RF atresia may be the continuation of the massive oocyte attrition occurring in fetal ovaries (4). It has been recently suggested that expression of the protooncogene receptor c-kit is essential for germ cell survival (Ref. 27 and references therein). Therefore, a defect in the W locus that encodes this protein may have occurred during the multiple successive divisions of germ cells during fetal life and could be responsible for the observed massive neonatal RF atresia. In adult rodents, it has been proposed that cyclic levels of gonadotropins may protect RF from atresia (21). However, in hypophysectomized mice and rats presenting undetectable levels of pituitary hormones, RF atresia is low (28, 29). Consequently, pituitary factors, including an unidentified atresia-promoting factor, have been speculated to play a direct or indirect role in RF atresia (28). Hence, we must admit that the mechanisms regulating the RF depletion by atresia in both juvenile and adult animals are not clearly identified and require further studies. B. Depletion of the stock by entry of follicles in growth phase: initiation of follicular growth 1. Morphometric studies. Follicular growth may begin at any time of the female's reproductive life, as early as infancy, and continues throughout puberty, during the ovarian cycle, and pregnancy, until the end of the reproductive period (15). The precise time at which follicular growth actually starts is difficult to establish. In humans, primordial, intermediary, and small primary follicles differ in their diameter, due to differences in the number and size of granulosa cells (GCs), but they do not differ in mean diameter of their oocyte and its nucleus (9). In addition, studies performed in rats have indicated an extremely slow proliferation rate of GCs from follicles containing between five and eight GCs in the largest cross-section (30). This illustrates the very slow process by which primordial follicles are transformed into small primary follicles, which is seen more as a maturation than a growing process. In humans, evolution of morphometric parameters of RF and early growing follicles shows that active follicular growth starts when the oocyte nucleus reaches a critical diameter of 19 jam and when approximately 15 GCs are present in the largest cross-section (9). This finding is in keeping with observations made in mice and rats where the entry into the growth phase takes place when eight to 20 GCs are present in the largest cross-section (31,32). In rodents, nucleoplasmic RNA syntheses may be the first event in the initiation of oocyte growth in response to an inductive stimulus coming from surrounding follicular cells (31, 33) or from the oocyte itself (32). 2. Factors triggering the initiation offollicular growth. The possible existence of a factor triggering the initiation of follicular growth remains an intriguing and unanswered question. a. Is there an inhibitory factor preventing RF from leaving the stock? In humans (13) as in other mammals, a strong correlation exists between the RF stock size and the number of growing follicles. The larger the stock, as in the infant ovary, the greater the number of follicles beginning to grow. With aging, as follicles continuously leave the RF stock, the number of growing follicles decreases, but the proportion of primary and early growing follicles increases in primates (9,34) as in mice (35). This increase may be triggered by the progressive disappearance of an inhibitory influence on primary follicle growth exerted by primordial and intermediary follicles themselves, acting via an unknown mechanism whose magnitude is directly proportional to the size of the stock. This putative mechanism may represent an adaptive process, allowing the ovary to promote the growth of a sufficient number of follicles to permit one of them to achieve its complete development and ultimately to ovulate, despite declining numbers of available follicles. However, the size of the RF stock does not appear to be the only factor controlling

4 124 GOUGEON Vol. 17, No. 2 ' the number of follicles starting to grow per time unit, since, during pregnancy in mice, this number decreases (36). b. Is there a stimulus acting on the RF store that stimulates some follicles to leave it? Data concerning a possible stimulatory role for gonadotropins in the initiation of growth phase in immature primates appear confusing. In contrast to the hypophysectomized monkey fetus in which initiation of follicular growth does not appear blocked (37), there are no early growing follicles in the anencephalic human fetus (38). Despite some studies suggesting that gonadotropins are involved in the initiation of follicular growth in immature rodents (24, 39, 40), they have been confirmed neither by observations of ovaries coming from mice bearing a natural deficiency of gonadotropins (20, 21) nor by transplantation of neonatal rat ovaries into recipient adult females exhibiting either high or low levels of circulating gonadotropins (26). In addition, the efficiency of gonadotropins in initiating RF growth between days 1 and 5 of life is in apparent contradiction with recent data on presence of gonadotropin receptors in RF. Although shortened receptors for FSH (FSH-R) can mediate a response to FSH (41), full-length transcripts for the FSH-R have not been detected in the mouse ovary until day 5 of life (42). FSH-R and LH receptor (LH-R) only appear in the ovary from days 5 to 11 (43), whereas immunostaining for FSH and LH to rat RF was shown only from day 8 of age (44). Thus, data from immature rodents suggest that initiation of follicular growth may be independent of a gonadotropic effect. During adult life in rodents, initiation of follicular growth may be controlled by gonadotropins (20, 21, 45). In humans, the acceleration in the decay rate of RF reported after 38 yr of age (13) may be linked to the significant increase in circulating FSH levels in women over 35 yr of age (46). Nevertheless, during natural hypopituitary conditions in both animal species and humans, the initiation process is not completely abolished (20, 21, 47). Two studies in mice (24) and rats (26) and more recent findings in monkeys, in which a GnRH agonist (GnRH-a) blocked the transition between intermediary and primary follicles (34), suggest that gonadotropins may act on the transformation of flat GCs into cuboidal GCs. Rather than initiating follicular growth, per se, gonadotropins may act on RF maturation by transforming their GCs, making it possible for these follicles to enter the growth phase in response to an unknown signal. Recent findings indicate that some of the genes coding for certain proteins, growth factors and their receptors (48-53), or protooncogenes (27), which are present in GCs and /or the oocyte of small follicles, may be involved in the transition of RF to growing follicles. For example, it has been suggested that pituitary hormones such as TSH may act synergistically with FSH to enhance the entry of RF in the growth phase (20). Thus, the erba oncogene, which encodes a thyroid hormone receptor, is expressed more abundantly in GCs from small follicles than in GCs from large follicles (54). It has also been proposed that expression of steel, encoding the Kit ligand and present in GCs, may be involved in stimulation of oocyte growth via c-kit receptor tyrosine kinase present on this cell (27), and that abundant expression of myc protein in the human primordial oocyte may play a role in the initial growth of the oocyte (55). Finally, in primordial and preantral follicles, the retinoblastoma gene product (prb) expression is different in GCs and in the oocyte (33). In primordial follicles, the nonproliferating flat GCs show low prb expression, but the oocyte exhibits strong nucleolar prb staining. It has been suggested that prb expression in the oocyte nucleolus is associated with production of negative growth ) factors that may inhibit proliferation of GCs. During resumption of follicular growth, GCs show increased prb staining, and the enlarging oocyte shows prb depletion. These data suggest that the resting state of the primordial follicle may be maintained by the oocyte, and resumption of follicular growth may be initiated by GCs. To overcome negative signaling from the resting oocyte, the GCs of human primordial * follicles may require stimulation by factors arising from vascular pericytes such as the Thy-1 differentiation protein (33). Although the number of follicles entering growth can be modulated either by the size of the pool or by endocrine factors, recent studies suggest multiple paracrine/autocrine factors acting in an elusive fashion. In conclusion, we must., admit that mechanisms triggering atresia of RF and initiation of follicular growth remain largely unknown, especially in primates. Consequently, identification of the chemical nature and cellular source of the putative inducer of follicular growth initiation remains a promising field of investigation. III. Morphological Aspects of Follicular Development In this section, the anatomical and dynamic aspects of the primate folliculogenesis will be described from the entry of a follicle into the growth phase until it becomes preovulatory. A. Description of the follicular development in primates When follicles enter the growth phase, they enlarge, both by proliferation of GCs and by an increase in size of the * oocyte. A zona pellucida begins to be laid down around the oocyte shortly after initiation of follicular growth; however, its exact cellular source remains unclear (56). Progressively, follicles become secondary follicles (Fig. 2A). At this stage of development some stroma cells near the basal lamina become aligned parallel to each another. As the follicle enlarges, this surrounding connective tissue stratifies and differentiates into two parts. The outer part, the theca externa, is composed of cells that do not differ in any respect from the cells of the undifferentiated theca. In the inner part, the theca interna (TI), some fibroblast-like precursor cells change into polyhedral cells, also referred to as epithelioid cells (Fig. 2B). Morphological studies have shown that definitive theca lay- 4 ers appear only when follicles contain three layers of GCs (follicle diameter: /nm) in monkeys (M. J. Koering, unpublished data), and three to six layers of GCs (follicle diameter: /um) in women (A. Gougeon, unpublished data). The smallest early growing follicles lack an independent blood supply, but secondary follicles /am in diameter are served by one or two arterioles, terminating in ' an anastomotic network just outside the basal lamina (57). The physiological importance of this event is emphasized by the fact that the follicle becomes directly exposed to factors circulating in the blood. From the time of appearance of epithelioid cells in the TI, the secondary follicle is defined as a preantral follicle (Fig. < 2C) and constitutes the first class of growing follicles in a

5 April, 1996 OVARIAN FOLLICULAR DEVELOPMENT IN PRIMATES 125 t FIG. 2. Human ovarian follicles with undifferentiated GCs. A, Secondary follicle, X275. B, High power micrograph of epithelioid TICs (into open circles) from a preantral follicle, X480. C, Preantral follicle (class 1), X350. GC, Granulosa cell; TI, theca interna; TE, theca externa; BL, basal lamina. [A-B reproduced with permission from A. Gougeon, The Ovary, pp (5).] classification based on morphological aspect and total num- * ber of GCs in each individual follicle (Fig. 3). Appearance of an antral cavity starts with the development of small fluid-filled cavities of 40 jam in diameter that aggregate to form the antrum (Fig. 4A). From this point, the GCs surrounding the oocyte constitute the "cumulus oophorus," In primates, as in most mammals,, shortly after the follicle acquires a single antral cavity, the GCs that border the basement membrane lose their cuboidal shape and assume a columnar appearance, while the remaining GCs remain cuboidal (4). In ewes, these columnar GCs exhibit a greater rate of cell proliferation than do GCs in the antral region (58). In humans and monkeys, follicles pass from the preantral (class 1) to the early antral stage (class 2) at a follicular diameter between 180 and 250 /xm (10, 59). Through the accumulation of fluid in the antral cavity and proliferation of GCs and TI cells (TICs), the follicle progresses through subsequent stages of development (Fig. 4B, class 3; and Fig. 4C, class 4) until it reaches a size of about 2 mm. Healthy follicles measuring 2-5 mm (class 5), referred to as selectable follicles, are observed at all stages of the cycle. Those that are present during the late luteal phase constitute the population from which the follicle destined for ovulation during the subsequent cycle will be selected (60, 61). Their number in the late luteal phase is between three and 11 per ovary in 24- to 33-yr-old women (62) and strongly decreases with increasing age. In macaca monkeys, follicles of about 1 mm constitute the population of selectable follicles (63). There is no healthy 5- to 9-mm follicle (class 6) during the midluteal phase, and they are rarely observed during the late follicular and luteal phases. During the early follicular phase, the newly selected follicle belongs to this class, and its diameter is between 5.5 and 8.2 mm (60). The size of the follicle destined for ovulation increases greatly during the follicular phase by cellular multiplication and accumulation of fluid in the antrum until the ovulatory gonadotropin surge, and then only by the latter process until ovulation. In humans, the diameter of the preovulatory follicle increases from 6.9 ± 0.5 mm (early follicular phase; range, mm) to 13.3 ±1.2 mm (midfollicular phase; range, mm) and then to 18.8 ± 0.5 mm (late follicular phase; range, mm) (60, 62). The mean number of GCs increases from approximately 2-5 million during the early follicular phase to million at the time of ovulation (59). During the final phase of development, the granulosa layer of the follicle destined to ovulate is subjected to marked morphological transformations (Figs. 5 and 6) resulting from a modulation of the actin cy toskeleton organization, which may be intrinsic in the programmed events that lead to GC differentiation (64). While mitotic figures have disappeared after the midcycle gonadotropin surge, as many as 1.2% and 0.3% pyknotic GCs can be observed in monkey (65) and human (66) preovulatory follicles, respectively. Color Doppler studies showed that there is a direct correlation between the development of human ovarian follicles and their blood flow (67). Therefore, from the midfollicular phase, the primate preovulatory follicle becomes a highly vascularized structure with the area of the TI occupied by blood vessels being 2 times larger than in other follicles within the same or contralateral ovary (68, 69). This may occur as a result of an active endothelial cell proliferation in the TI throughout the follicular phase (70). Some hours before ovulation, the granulosa wall, which was avascular before the midcycle gonadotropin surge, appears to be bloody after invasion of blood vessels originating from the TI and occurring after the surge (70).

6 126 GOUGEON Vol. 17, No. 2, 375x10 3 GC r ecru i table / 1.87x10 6 GC FIG. 3. Classification of follicles in the human ovary. early preovulatory 47x10 6 GC " 01 early growing ^INITIATION 0.04 mm x _* \ resting follicle OVULATION B. Cytological and morphometric changes offollicular tissues during development During follicular development, the granulosa layer undergoes numerous changes (Table 1). Independently of its cyclic variations, the rate of cell proliferation increases as the follicle develops (71) and culminates in selectable follicles observed during the late luteal phase and in the newly selected follicle. The mean number of GCs per cubic millimeter of tissue decreases significantly as follicles mature. This corresponds to an increase in volume of individual GCs and can be considered as morphological evidence of GC maturation (66). Ultrastructural studies show that immature GCs contain mainly rough endoplasmic reticulum, which reflects the preponderance of protein synthesis for growth and proliferation. A gradual increase in the extent of smooth endoplasmic reticulum and appearance of mitochondria with tubular cristae and lipid droplets occur as steroidogenesis increases in GCs from preovulatory follicles and culminate with maximal progesterone (P) secretion in the corpus luteum (64). GCs communicate between themselves via two types of specialized cell contact systems. Gap junctions are aggregates of transmembrane channels, each composed of six identical subunits, or connexins, which are arranged to form a central pore (72). Adherent junctions are associations of filamentous dense material on the intercellular sides of the plasma membrane. Adherent junctions are present as early as the first stages of follicular growth whereas gap junctions appear during antral cavity formation and grow rapidly during expansion of the antrum. In GCs from human preovulatory follicles, approximately 30% of the entire plasma membrane is occupied by gap junctions. Whereas FSH increases acquisition of gap junctions by GCs, the gonadotropin surge or human chorionic gonadotropin (hcg) administration to preovulatory GCs causes a significant decline in gap junctions (64). There are few data concerning morphological changes of TICs during follicular growth. Through mitosis of theca stem cells, their number increases, but to a lesser extent than do GCs. C. Dynamics of human follicular growth In the human, the time for early growing follicles to attain the preantral stage is not known. From data obtained in monkeys (73), where less than 10% of follicles possessing one to two layers of GCs contain GCs labeled with [ 3 H] thy midine, and rats (30), where the growth rate of small follicles is very slow, it can be hypothesized that this time is probably more than several months in humans. Throughout development, the number of GCs increases in each individual follicle. Consequently, each follicle enters a given class when its number of GCs attains the lower limit of this class (characterized by a fixed number of GCs: Fig. 3). After a time lag (the transit time) directly proportional to its GC mitotic index, the follicle leaves this class when its number of GCs attains the upper limit of the class and enters into the subsequent class. Taking into account that from the preantral stage (class 1), the ovulatory follicle will pass through eight classes to reach the preovulatory size (class 8), the time required for this follicle to pass from the preantral to the ovulatory stage will be the sum of the transit times required

7 April, 1996 OVARIAN FOLLICULAR DEVELOPMENT IN PRIMATES 127 simultaneously with an entry of follicles into class x + 1. Visible passages of follicles from one class to the next were observed during the late follicular phase when class 1 follicles enter class 2, between the late luteal and early follicular phases when class 2 follicles enter class 3, during the late follicular phase when class 3 follicles enter class 4, and during the late luteal phase when class 4 follicles enter class 5 (71). The second technique calculated the GC doubling time after in vitro determination of the duration of mitosis and estimated the mean mitotic index for each follicular type (Table 2). The duration of mitosis obtained, 1.42 h (71), was similar to that obtained in the ewe (74). From these transit time values, going back from the preovulatory follicle to preantral follicles has suggested the dynamics of human follicular development (Fig. 7). Functional experiments studying the effects of follicle-ectomy or luteectomy, both in monkeys (75) and humans (76, 77), are in agreement with the dynamics of follicular growth, at least from the selectable stage onward., ~,» FIG. 4. Human ovarian follicles with undifferentiated GCs. A, Early antral follicle (class 2), X170. B, Small antral follicle (class 3), X60. C, Small antral follicle (class 4), X25. [A and C reproduced with ' permission from A. Gougeon, The Ovary, pp (5).] i for this follicle to pass each class. These transit times have been calculated by two different techniques. The first measured the time interval separating the visible entry of follicles into one class (the number of healthy follicles increases and their mean size decreases, signifying that small follicles have entered the class) from the visible departure (when the population of follicles decreases) of these follicles to the subsequent class (71). These changes have a physiological significance when a departure of follicles from class x occurs IV. Gonadotropins and Follicular Functions A. The primary role of gonadotropins to initiate and sustain follicular functions The basic function of the ovulatory follicle is to produce a fertilizable oocyte and to function as an endocrine gland during its maturation and after ovulation through transformation into a functional corpus luteum. Preovulatory follicular cells, and especially GCs, become able to assume their full endocrine functions at the completion of a highly regulated process that leads them from a proliferative to a differentiated status. This process is under the primary control of two pituitary hormones, FSH and LH. 1. Gonadotropins sustain follicular growth. Whereas the role of gonadotropins in initiating follicular growth remains unclear, many studies have demonstrated their tropic action on follicle cells. In rodents, development of GCs is impaired in primary and secondary follicles when levels of circulating gonadotropins are drastically reduced either naturally or experimentally (21, 26, 78). In these situations, GC nuclei are irregularly shaped rather than round, the spaces between the GCs disappear, and GC proliferation is strongly impaired but not entirely inhibited. In all cases, injection of FSH restores the regular round shape of GC and development of small follicles (78, 79). That gonadotropins and especially FSH are of primary importance to sustain follicular growth has been extensively reviewed (4, 6). Several examples illustrate this primary role of gonadotropins in humans. Various physiological and pathological situations are characterized by an unfavorable hormonal environment either due to low levels of circulating gonadotropins or to an unfavorable FSH/LH ratio, associated with anovulation. Women suffering from Kallman's syndrome (80) can be induced to ovulate with exogenous gonadotropins, and those suffering from a poly cystic ovarian syndrome (PCOS) in which circulating levels of FSH are very low, can be induced to ovulate by treatment with high doses of FSH (81, 82). In normal women undergoing in vitro fer-

8 128 GOUGEON Vol. 17, No. 2 FIG. 5. Follicular wall changes in the follicle selected for ovulation (X 480) during the early (A) and midfollicular phase (B). The granulosa layer is more or less homogeneous, with some GCs that border the basal lamina assuming a columnar appearance, and polyhedral GCs exhibit a high proliferative activity. C, During the plasma E 2 peak, all the GCs that border the basal lamina assume a columnar appearance with nuclei ranked along the basal lamina, whereas other GCs remain polyhedral in shape with a declining mitotic activity. GC, Granulosa cell; TI, theca interna; TE, theca externa; mitosis are in open circles. FIG. 6. Follicular wall changes in the preovulatory follicle from the midcycle gonadotropin surge (X480). A, When the preovulatory gonadotropin surge begins, the columnar appearance of the basal layer of GCs disappears, the basal lamina becomes blurred, and mitoses become scarce; the size of TICs is maximum. B, Just after the midcycle gonadotropin surge, the basal lamina is disappearing and GCs are homogeneously distributed, there are no more mitosis, and the size of TICs has decreased. C, A few hours before ovulation there are no more mitoses, GCs are dissociated, spindle-shaped, and radially arranged, and the granulosa wall appears to be bloody after invasion of blood vessels coming from the TI. GC, Granulosa cell; TI, theca interna; BL, basal lamina. >* * c > tilization and embryo transfer (IVF&ET), treatment with exogenous gonadotropins overrides the normal ovulatory quota by sustaining preovulatory growth of a large number of follicles (83). Moreover, recent data suggest an exclusive role for FSH in sustaining follicular development. In monkeys treated either with high doses of GnRH antagonist (84) or with a LH antiserum (85), leading to LH-deficient animals, FSH alone is capable of stimulating ovarian follicular growth. 2. Gonadotropins sustain follicular steroidogenesis. According to the "two-cell, two gonadotropins" theory (86), TICs are stimulated by LH to produce aromatizable androgens that are

9 " April, 1996 OVARIAN FOLLICULAR DEVELOPMENT IN PRIMATES 129 TABLE 1. Morphometric characteristics of GCs in healthy growing follicles (numbers in parentheses) during folliculogenesis Class (fim) Number of GC Mitotic index (%) 1(742) 2(565) 3 (342) 4(143) 5(133) 6(31) 7(16) 8 (22) 8 (16) 6 2,100 it 0,100 7,600 d: 0,100 33,300 i: 1, ,800 i: 7, ,500 Ht 44,900 3,400,000 it 300,000 24,140,000 dt 2,970,000 54,760,000 it 4,610,000 59,070,000 ± 5,510, t it jt it it ^t dt dt dt Values are expressed as mean ± SEM. Before the midcycle gonadotropin surge. 6 After the midcycle gonadotropin surge. c Measurement not possible because of the absence of an antral cavity. GC (X 1(T 5 ) per mm 3 of tissue ± ± ± ± ± ± ± ± ± 0.20 Thickness of GC layer c 35.9 dt it ^t it it ; t it = t3.2 TABLE 2. Transit time for each follicle class; comparison of results obtained from in vitro studies and analysis of morphometrical parameter changes Follicle class Mitotic index 0.38 ± ± ± ± ± « «2.38 c Doubling time (h) 246 ± ± ±8 111 ± 6 60 ± « «115.6 Values are expressed as mean ± SEM. During the late luteal phase. 6 Range of individual GC mitotic indices observed during the early follicular phase. c Range of individual GC mitotic indices observed during the midfollicular phase. From doubling time value 23.8 ± ± ± ± ± << «11.2 Transit time (days) From time interval between entry and exit transported to GCs where they are converted to estrogens by aromatizing enzymes, which are induced by FSH. The human thecal steroidogenesis and its control by LH have been extensively documented (84, 87, 88), and the recent use of recombinant LH has confirmed previous observations (89). Briefly, the TI appears to be the primary site of androstenedione (A) synthesis, which is the principal aromatizable steroid, whereas testosterone (T) is produced in lesser amounts (90). The androgen secretion by TICs under LH stimulation appears to result from activity of enzymes, such as cholesterol side chain cleavage (P450 scc ), 17a-hydroxylase/lyase (P450 17a/lyase ), and 3)3- hydroxysteroid dehydrogenase (3/3HSD). The steroidogenic pathways in GCs are organized principally for the metabolism of androgens to estrogens before the midcycle gonadotropin surge and for de novo synthesis of P and its metabolites after the midcycle LH surge (91). Briefly, 17/3-hydroxysteroid dehydrogenase (17HSD), which is present in GCs as soon as the primary stage (92) transforms A into T, and aromatase (also referred as P450 arom ) transforms T into E 2 (88). Although aromatase has been immunohistochemically detected in TICs of follicles from the preantral stage onward (93), it has been estimated that during the preovulatory maturation, 99% of the aromatase activity resides in the GCs, which have at least 700 times the aromatase activity of the TI (94). The role of FSH in the induction of GC aromatase activity has been demonstrated in rats in vivo by using recombinant human FSH (95). During the late preovulatory maturation, acquisition of LH receptivity transforms the ability of GCs to respond to gonadotropins since aromatase can be stimulated by hcg (96). After the midcycle gonadotropin surge, 3/3HSD is expressed, and GCs turn from a predominantly estrogen-producing to a progestinproducing tissue (66). Data reported above indicate that gonadotropins have long been recognized as basic supports for follicular growth and steroidogenesis. However, in vivo analyses of folliculogenesis show that small and large follicles do not respond in a similar way to the same gonadotropic signal. B. Changes in follicular responsiveness to gonadotropins during folliculogenesis 1. The basal follicular growth. a. Morphological evidence supporting the existence of a basal follicular growth. Throughout human infancy, when the circulating levels of gonadotropins are lower than those recorded during puberty and adulthood, 2- to 6-mm follicles are observed in the ovaries, and some of them can exceed 6 mm (97) when circulating gonadotropin levels increase after 6 yr of age (98). During pregnancy, when plasma FSH levels are low or undetectable (99), the human ovary contains many small healthy and atretic antral follicles with diameters up to 6 mm (100). In patients suffering from Kallman's syndrome, levels of gonadotropins are very low (101). In these women, follicular development is markedly impaired (102); nevertheless, the observation of atretic antral follicles (47) and the

10 130 GOUGEON Vol. 17, No " BASAL FOLLICULAR GROWTH 1st cycle LL 3rd cycle FIG. 7. Dynamics of the development of follicles belonging to the cohort from which the follicle destined to ovulate will be selected. Growth commences with the entry of follicles into class 1 during the early luteal phase (EL). Twenty-five days later, during the late follicular phase (LF) of the following cycle, an entry of follicles into class 2 can be seen. The end of the follicular phase is a favorable period for the appearance ^ of the antrum since the circulating levels of FSH have been high during the preceding days, and this hormone promotes the initiation of antrum formation. Twenty days later, follicles pass into class 3 during the late luteal phase (LL) and 15 days later into class 4 during the late follicular phase (LF) of the subsequent cycle. Follicles enter into class 5 (selectable stage) 10 days later, in the late luteal phase (LL), and constitute a population from which the follicle destined to ovulate during the subsequent cycle will be selected. [Adapted with permission from A. Gougeon, The Ovary, pp (5).] occurrence of ovulation within approximately 2 weeks after exogenous gonadotropin treatment (80, 82) suggest the presence of selectable follicles at the onset of treatment. In oligomenorrheic women, and in patients with polycystic ovaries showing a very low FSH/LH ratio (103), follicles can attain more than 5 mm. In the so-called "resistant ovary syndrome," women presenting with primary amenorrhea are usually hypoestrogenic despite high serum levels of FSH and LH. Ovarian biopsies from these women show a progression of follicular growth up to 1-2 mm (104). Taken together, these findings indicate that in an inadequate gonadotropic environment, human folliculogenesis is impaired since ovulation does not take place. However, although the number of growing follicles in these conditions is more or less markedly reduced, follicular growth is not completely abolished since in the less favorable conditions, some follicles can grow to a diameter of approximately 2-5 mm, i.e., the selectable stage. So, according to some authors (103), the growth of follicles from the preantral to the selectable stage only requires tonic (or basal) levels of gonadotropins. This could explain why healthy selectable follicles 2-5 mm in diameter are always present from infancy to menopause in humans, even when plasma levels of gonadotropins are low. b. Low follicular responsiveness to gonadotropins during basal growth. Although the rate of GC proliferation increases as follicles enlarge (Table 1), it is low and not related to cyclic hormonal changes for follicles smaller than approximately 2 mm (classes 1-4) (105). In addition, these follicles seem to be unresponsive to exogenous gonadotropins. In women undergoing IVF&ET, the mean number of recovered oocytes after treatment with human menopausal gonadotropins (hmg) is grossly similar (106) to the number of selectable follicles (>2 mm) that are* present during both late luteal and early follicular phases of the spontaneous cycle (60, 62, 107). During basal growth, primate follicles exhibit slight steroidogenic activity. Small and large human preantral follicles cultured in vitro spontaneously produce very low amounts of P and A whereas E 2 production is not detectable (108). Cultured GCs from marmoset follicles mm (109) and from human follicles smaller than 2 mm (110) possess only faint aromatizing capacity. In human growing follicles, immunolocalization of steroidogenic enzymes shows that neither - P450 17a nor P450 scc are present in TI from follicles smaller than 2 mm (111, 112). Immunoreactive 3/3HSD has been detected in both TICs and GCs of preantral follicles (92); j however, the staining is slight and possibly not significant. This observation is in contradiction with previous immunohistochemical studies showing that 3/3HSD is restricted to TICs in follicles larger than 2 mm (113). Similarly, the presence of P450 arom in follicles during basal growth is disputed. This enzyme has been reported to be either absent (112) or present (93) in GCs of these follicles, despite the same aro- * matase monoclonal antibody being used. These discordant results may be due to differences in the processing of the samples [either fixed paraffin section (93) or unfixed frozen sections (112)]. It can therefore be assumed that during basal growth, the development of human follicles requires only small levels of ^ gonadotropins. These follicles are faintly responsive to cyclic

11 April, 1996 OVARIAN FOLLICULAR DEVELOPMENT IN PRIMATES 131 gonadotropin changes, and their GCs exhibit low proliferative activity, weakly stimulated by gonadotropins. In addition, except for 17HSD (92), steroidogenic enzymes are weakly or not expressed. ^ 2. The selectable follicles. From a size of 2 mm, follicles become more dependent on FSH since their percentage of atresia decreases when FSH increases (105). From the mid- to the late luteal phase, granulosa of selectable follicles exhibit a significant increase in the rate of cell proliferation (105), paralleling the increase in circulating levels of FSH. In addition, Hwhen stimulated with hmg during both the late luteal and early follicular phase, selectable follicles can grow more < quickly than unstimulated selectable follicles (114). Observations made at various times during the menstrual cycle indicate that the intrafollicular steroid milieu from healthy and atretic 1- to 5-mm follicles is similar (Table 3). These follicles have a low intrafollicular estrogen-androgen ^ ratio and A is the dominant steroid. Immunohistochemical studies have shown that both 3/3HSD (113) and P450 arom (93) are present in TICs from selectable follicles, but that both P450 scc and P450 17a are absent (111, 112). The presence of A in follicular fluid (FF), however, suggests that these latter enzymes are probably insufficiently expressed to be detected by immunohistochemical methods. Although the percent conversion rate of A to E 2 by cultured GCs increases slowly when the follicle diameter exceeds approximately 3 mm (110), P450 arom has not been detected in most of the studies (111, 112, 118, 119). These observations, therefore, indicate that selectable follicles become more responsive to gonadotropins in terms of quality and growth rate but that their FSH-induced aromatase remains poorly expressed.» 3. The neiuly selected follicle. During the early follicular phase, just after having been selected, the follicle destined for ovulation contains detectable levels of FSH (120), and the mitotic index of its GCs is very high (60). Whereas A was largely predominant in FF of selectable follicles, its concentration decreases in newly selected follicles while the E 2 concentration strongly increases as compared with selectable follicles, * leading to a low FF A/E 2 ratio (Table 3). This is due to P450 arom that is expressed in GCs from follicles of a similar size (111). These data, therefore, indicate that during selection, the follicle destined for ovulation changes from an androgen- to an estrogen-producing structure by expressing its FSH-induced aromatase activity. 4. Maturation of the preovulatory follicle. In parallel with its increasing size, high proliferative activity of GCs, and morphological alterations of the theca and granulosa layers, the preovulatory follicle undergoes marked changes in steroidogenic activity as its maturation progresses. Enzymes such as 3/3HSD, P450 17a and P450 scc are detected at a significant level in TICs from these follicles ( ,121), and P450 arom enzyme exhibits strong staining in GCs when the follicular diameter is more than 11 mm (93,112). Significant levels of mrnas, for P450 scc and P450 17a appear in TICs of preovulatory follicles from D7 and D8, respectively (119). The levels of mrnas for P450 arom, which were low on D8 and D9, strongly increase on D13 (119). Consequently, steroid production increases strongly, and from the early to late follicular phase before the midcycle gonadotropin surge, E 2 in FF increases from around 658 to approximately 2,396 ng/ml, and mean concentration of steroids increases from 2,275 to 4,829 ng/ml (Table 3). These studies are in keeping with early biochemical studies indicating that aromatase activity increases progressively during the preovulatory phase, reaching 200 times that found in cells obtained from selectable follicles (94). Thus, during the follicular phase, GCs from the selected follicle exhibit an increasing aromatase activity. Since circulating levels of FSH are high at this time, it can be assumed that GCs of the early preovulatory follicle become fully responsive to FSH. The full responsiveness of preovulatory GCs to FSH is also confirmed by progressive acquisition during preovulatory maturation of other FSH-induced functions such as the expression of LH-R, which is induced under the primary control of FSH in most species (122). During preovulatory maturation a dramatic increase in LH binding occurs in human GCs (123, 124). Once they have synthesized LH-R, preovulatory GCs become able to respond to the midcycle gonadotropin surge. During the late follicular phase, after the gonadotropin surge, the mean concentration of steroids in FF increases from 4,829 to 10,995 ng/ml (Table 3). From the time of the gonadotropin surge, GC proliferation becomes extremely low and FF concentrations of E 2 and A decrease, whereas those of P and 17o:-hydroxyprogesterone (17a-OHP) increase sharply (Table 3). Concomitantly, 3/3HSD detected both by enzymohistochemistry (121) and immunolocalization (113) exhibits strong activity in GCs. Thus, when folliculogenesis is completed, just before ovulation, GCs are highly differentiated in the preovulatory follicle, having stopped to proliferate but producing high levels of steroids. In conclusion, the analysis of folliculogenesis indicates that follicular responsiveness to gonadotropins increases progressively as the follicle develops from the preantral to preovulatory stage. How this occurs in primates remains poorly TABLE 3. Changes in FF mean (±SEM) concentrations (ng/ml) of steroids from the selectable to the preovulatory stage Follicle type Atretic (1-5 mm) Selectable Newly selected (EF) Preovulatory (MF) Preovulatory (LF:E 2 peak) Preovulatory (LF:E 2 <<LH surge) Preovulatory (after the LH surge) E 2 20 ±5 15 ±5 658 ± 38 1,270 ± 161 2,396 ± 348 2,583 ± 228 1,109 ± 142 A + T + DHT 794 ± ± ± ± ± ± ±21 17a-0HP 713 ± ±112 1,002 ± 212 1,812 ± 142 2,034 ± 326 P 73 ± ± ± ± 74 1,228 ± 228 2,464 ± 226 7,773 ± 643 Total ,275 2,712 4,829 6,146 10,995 Mean concentrations are estimated from Refs 61, EF, Early follicular phase; MF, midfollicular phase; LF, late follicular phase.

12 132 GOUGEON Vol. 17, No. 2 understood but suggests specialized modifications at the cellular level. The wealth of information existing on ovarian responsiveness obtained through in vitro and in vivo studies may provide clues as to the nature of the intraovarian events responsible for the increase in sensitivity to gonadotropins. We will review the existing data in the following section. V. Factors Regulating the Progressive Acquisition of Follicular Responsiveness to Gonadotropins A major question in this issue may be: is the acquisition of follicular responsiveness to gonadotropins solely attributable to changes in gonadotropin receptor content or function, or is it also attributable to molecules acting at the level of follicular cells? A. The gonadotropin receptors Gonadotropins act on follicular cells by coupling with receptors localized at the cell surface. FSH binds exclusively to FSH-R in GCs, whereas LH binds to LH-R, initially localized to TICs, and then to GCs during maturation of the preovulatory follicle (125). FSH-Rs have been detected on GCs of human follicles from the secondary stage onward (123, 124). Despite the absence of data on the precise stages at which FSH-Rs increase in number, the density of FSH binding to GCs is higher in antral than preantral follicles and does not vary when antral follicles grow from 3-12 mm (124). However, human preovulatory GCs in the late follicular phase bind less FSH than preovulatory GCs in the midfollicular phase (124). Despite the synthesis of FSH-R on GCs being modulated by several factors, the process has long been recognized to be under the primary control of FSH (126). Yet, a recent study indicated that in kpg mice lacking circulating gonadotropins, the ovaries contain FSH-R mrnas, which suggests that factors other than FSH act on the expression of the FSH-R gene (42). Nonovulatory follicles in both monkeys (127) and humans (128) do not possess LH-R on their GCs since hcg does not stimulate their steroidogenesis. Receptors for LH appear in preovulatory GCs starting in the midfollicular phase (124, 127) and are in highest density in mural GCs located next to the basal lamina (124). TICs possess LH-R at all stages from the preantral stage onward as demonstrated by autoradiographic and immunocytochemical studies in monkeys (69) and humans (123, 124). The LH binding to primate TICs starts with the appearance of epithelioid cells and increases during follicular maturation. Binding is more important in TICs of 3-mm follicles than in preantral or early antral follicles and is stronger in 4- to 7-mm as compared with 2- to 3-mm follicles (124). Taken together, these data indicate that the number of LH-Rs increases in TL and that of FSH-Rs increases first, then decreases as the primate follicle develops. Nevertheless, data concerning the number of FSH-Rs on GCs remain imprecise, especially at times when an increasing follicular responsiveness to gonadotropins has been detected. Further studies are therefore required to analyze the dynamic changes of GC content in FSH-Rs and their mrnas at each critical stage of follicular development. However, one question must be addressed: do changes in the number of gonadotropin receptors or in the levels of their mrnas account for the changes of follicular responsiveness to gonadotropins during follicular growth? In some cases, the detection of mrnas for gonadotropin receptors cannot explain these changes. Because there are multiple transcripts and splice variants encoding the LH-R and FSH-R, it is possible that some of them do not exhibit functional activity (Ref. 122 and references therein). Hence, it has been hypothesized that increasing levels of FSH-R mrnas without an increase in active membrane receptors (129) may be caused by the transcription of an inactive splice variant and not the expression of mrnas encoding the func- * tional receptors (122, 129). Added to the complexity of the expression of the LH-R and FSH-R genes and multiple forms of mrnas are the transducing mechanisms to which these receptors couple. In rats, GCs isolated from preovulatory follicles exhibit greater FSH-responsive adenylyl cyclase than GCs isolated from small antral fol-. licles in which steroidogenesis is low (130), despite the similar numbers of FSH-binding sites in the two GC populations (131). In porcine GCs, FSH enhances steroidogenesis at concentrations that cause a decrease in the number of FSH-Rs, thereby suggesting either that only a small number of FSH-Rs are required to elicit a response in target cells or that less FSH is needed once the biosynthetic pathway is activated (129). These» observations have raised the possibility that there are different transducers to which FSH-R can couple or that adenylyl cyclase activity can be modified. Recent studies suggest that FSH-R and LH-R present at the early stages of follicular development may couple to subsets of transducers different from those to which they couple at later stages, which could lead to different physiological effects (122). Mechanisms regulating the gonadotropin receptor gene i expression and the transducing mechanisms to which these receptors couple are highly complex and unknown in primate follicle cells. Consequently, the precise role that gonadotropin receptor may play to increase the gonadotropin responsiveness of follicular cells during human folliculogenesis remains unelucidated and is a promising field of investigation. ' Irrespective of possible changes in levels or function of FSH-R, many experiments performed in mammals suggest that FSH action on GCs can be modulated by biological substances present in serum or synthesized by the follicle itself. Although FSH is required to promote normal follicular development, no growth-promoting effect of FSH has been ^ found on GCs from either small or large follicles cultured in vitro without serum in rats (132) and cattle (133). But, in the presence of serum (133, 134), FSH is an effective stimulator of cell proliferation. Conversely, GC differentiation only occurs in absence of serum (135, 136). These data, therefore, suggest the existence of factors in serum inhibiting GC differentiation and promoting GC proliferation. ) On the other hand, FSH acts on preantral human follicles cultured in vitro without serum to stimulate the proliferation of GCs, to induce the differentiation of an antral cavity, and to stimulate the production of E 2, P, and A (108). Similarly, FSH stimulates DNA synthesis of GCs in incubated whole follicles in hamsters (137) and mice (138) without serum or supplemented milieu, suggesting that follicular tissues, GCs

13 April, 1996 OVARIAN FOLLICULAR DEVELOPMENT IN PRIMATES 133 and/or TICs, produce substances acting in an autocrine or paracrine fashion, alone or in synergy with FSH to stimulate GC proliferation and differentiation. * Altogether, these observations lead to the hypothesis that in vivo FSH-induced functions in growing follicles may be mediated by factors coming from the serum or from the follicle itself, acting alone or in synergy with FSH. J3. The intraovarian peptides f There is evidence that the effect of gonadotropins on fol- - licular cell proliferation and cytodifferentiation is modulated by nonsteroidal products secreted by GCs and /or TICs. Growth factors have been recognized as peptides that may be involved in cell-cell interactions in the follicle., 1. The insulin-like growth factor (IGF) system. The IGF system consists of IGF-I and IGF-II, type I (IGF-I-R) and type II * (IGF-II-R) receptors, and a family of binding proteins (IGFBPs) that modulate the actions of IGFs (for review see Ref. 3). a. The IGF system in humans. In humans, IGFs alone possess stimulatory effects on both proliferation and steroidogenesis " in GCs and TICs from immature to preovulatory follicles (Table 4), and these effects are strongly enhanced by the presence of gonadotropins. The presence of IGFs and their receptors, as well as their respective mrnas, has been reported in human follicles at successive stages of their development (Table 5). Data are sometimes confusing. Whereas IGF-I and its mrna have not been detected in GCs, their presence in TICs is disputed.» IGF-I has been transiently detected in TICs from antral nonovulatory follicles by Hernandez et al. (160), but its mrna * has not been found by Zhou and Bondy (162). Conversely, El Roeiy et al (161) found IGF-I mrna but not IGF-I in TICs from follicles of similar size. There are several explanations to account for this disparity between gene expression and the absence of the gene product, including rapid secretion or the TABLE 4. Reported effects of IGFs on human follicle cells Granulosa IGF-I, alone or in synergy with FSH, stimulates proliferation of immature GC, mature GC, and IVF-GC ( ) IGF-I, alone or in synergy with FSH, stimulates aromatase mrna \ levels and aromatase activity in immature GC, mature GC, and IVF-GC (141, ) t IGF-I alone stimulates P production, independently of cell maturity, when GC are cultured in presence of calf serum (141, 144). In serum free conditions IGF-I alone is without effect (147) IGF-I enhances FSH- and/or LH-induced P production independently of cell maturity (141, 144, 147) IGF-II alone stimulates both E 2 and P production by IVF-GC when M they are cultured in presence of calf serum (148) > Theca interna IGF-I stimulates TIC proliferation in culture (149) IGF-I strongly enhances camp-stimulated production of 17a-0H P and P450 17a (150) IGF-I and IGF-II, alone or in synergy with LH and inhibin, strongly stimulate A production by TIC (89, 151) ^ In synergy with LH, IGF-I enhance nearly 4-fold A secretion and on average 2.4 times T secretion by TIC (151) * IGF-II, alone or in synergy with LH, stimulates A production (89) inability of the antiserum used to detect the small amount of protein present. Similarly, the production of IGF-II and its mrna by nonovulatory follicles is unclear. El Roeiy et al. (161) have detected IGF-II and its mrna only in preovulatory follicles, thus confirming the previous observation that cultured GCs sampled in women undergoing IVF & ET (IVF-GCs) can synthesize IGF-II (164). However, Zhou and Bondy (162), who have not studied preovulatory follicles, found IGF-II mrna in GCs from small growing follicles. Data reported in Table 5 suggest that the human follicle is a site of IGF-II rather than of IGF-I production. The situation prevailing in the human ovary appears to be different from that of the rat ovary, in which IGF-I gene expression is restricted to GCs (165). Under normal physiological conditions, the IGFs in body fluid are bound to a family of homologous proteins termed the IGF-binding proteins (IGFBPs) (for review see Ref. 3). Six different IGFBPs have been detected in the human ovary. IGFBP-I, -2, -3, and -4 have been detected in the fluid of follicles from 3 mm in diameter to the preovulatory stage (146,153,157, ), and IGFBP-6 has been found in FF of stimulated preovulatory follicles (173). According to the stage of follicular development, GCs contain mrnas encoding IGFBP-1, -2, -3, -4, and -5 (162,168,174) and cultured IVF-GCs synthesize IGFBP-1, -2, -3, and -4 (168, ). At all stages of follicular development, IGFBP-2 mrna levels are high in TICs (161), which also secrete IGFBP-3 in culture (175). There are consistent changes in levels of certain IGFBPs in association with the stage of follicle development. IGFBP-1 appears at higher levels in follicles after LH/hCG (167) and have been associated with luteinization of GCs ( ), and levels of both IGFBP-2 and -4 strongly decrease in preovulatory follicles ( ). Although a local amplifying effect of IGFBPs on IGF action has been suggested (179), it is more generally admitted that IGFBPs remove free IGF from follicular environment and, by preventing its binding to receptors, restrict IGF action on ovarian function (153). In humans, IGFBP-1 inhibits IGF-Imediated proliferation of IVF-GCs (140), and both IGFBP-1 and IGFBP-3 inhibit IGF-I-stimulated E 2 and P productions by these GCs (180). Moreover, in rats, all substantial inhibitions of FSH action induced by IGFBPs can be mimicked by anti-igf-i antibody (181). b. The GH/IGF axis. GH has recently been viewed as a "co-gonadotropin" capable of enhancing gonadotropin action at the ovarian level (182). GH and GH-releasing hormone (GHRH) have been reported to improve the effects of treatments with gonadotropins in patients undergoing IVF & ET whether normal or poor responders ( ), in women with anovulation and amenorrhea (186) or suffering from a PCOS (183), and in hypogonadotropic subjects (187). However, these observations are disputed. Other studies have shown no beneficial effect of GH administered to poor or normal responder IVF patients treated with similar doses of hmg ( ), and GHRH injected from cycle day 2 to the time of ovulation does not increase the number of fertilized oocytes and resulting pregnancies (191). In addition, in an infertile patient suffering from a GH deficiency (Oliver- McFarlane syndrome), adjuvant GH did not influence the ovarian response to exogenous gonadotropins (192).

14 134 GOUGEON Vol. 17, No. 2 > TABLE 5. Follicular levels of IGF-I, IGF-II, IGF-I-R, IGF-II-R, and their respective mrna (in parentheses) in human follicles Stage of follicle growth Primary Secondary Preantral (class 1) Small antral (classes 2-4) Selectable (class 5) Preovulatory (classes 6-8") Preovulatory (class 8 ) Stage of follicle growth Primary Secondary Preantral (class 1) Small antral (classes 2-4) Selectable (class 5) Preovulatory (classes 6-8 a ) Preovulatory (class 8 6 ) IGF-I - (-) IGF-I IGF-II + (+ ++) IGF-II Granulosa IGF-I-R +/++ (++) Theca interna IGF-I-R IGF-II-R + (+ + +) IGF-II-R + ( + ) Stage of follicle growth Selectable (classes 5 and 6) Preovulatory (classes 6-8 ) Preovulatory (class 8 6 ) Stage of follicle growth Primary Secondary Preantral (class 1) Small antral (classes 2-4) Selectable (classes 5 and 6) Preovulatory (classes 6-8 ) Preovulatory (class 8 6 ) IGF-I IGF-I Follicular fluid IGF-II : : ; IGF-II Oocyte IGF-I-R The staining or the level in FF was determined as positive (+ to ) or negative (-). a Selected follicle before the midcycle gonadotropin surge. 6 Preovulatory follicle after the spontaneous midcycle gonadotropin surge or after an hyperstimulation treatment. [Data from Refs. 48, 148, and ]. IGF-II-R If it exists, the nature of GH action at the ovarian level remains unclear despite GH being the primary regulator of serum IGF-I levels (3). In women treated with hmg and GH, IGF-I concentrations rise in FF (183) and serum (186) and are significantly higher in GH-treated women as compared with controls. However, IGF-I FF levels are significantly lower than serum levels (183, 186), and since GH is incapable of altering IGF-I and IGF-II mrnas in cultured IVF-GCs (158), the source of increasing circulating IGF-I is presumed to be the liver. Thus, GH action might be indirect by stimulating hepatic IGF-I, which can act alone or improve gonadotropin action at the follicular level. Alternatively, GH may exert a direct effect on GCs since GH receptors and their mrnas have been identified in GCs from nondominant and / or dominant follicles in natural or stimulated cycles (182, ). Taken together, the available data indicate that uncertainties remain regarding the role of GH in enhancing follicular growth. Spontaneous and induced pregnancies have been reported in two women with Laron-type dwarfism, a condition characterized by a deficiency of GH receptors and subsequent systemic IGF deficiency (196,197). This suggests that while circulating IGF-I and/or GH may contribute, their role is not indispensable for normal folliculogenesis in humans. In conclusion, available data indicate that IGFs may play a role in regulating both proliferation and steroidogenesis of human follicles by amplifying the stimulatory signals of FSH and LH in growing and preovulatory follicles. This impact of IGFs on both GCs and TICs argue in favor of the ovarian IGF-II system playing a key role in human ovarian follicular growth, whereas the role of IGF-I in regulating human follicular function is as an endocrine hepatic hormone rather than a paracrine/autocrine factor of follicular origin. 2. Epidermal growth factor (EGF) and transforming growth factor-a (TGFa). EGF and TGFa are polypeptides that have structural homology and bind to a common EGF/TGFa receptor (EGF-R) (for review see Ref. 198). In humans, EGF stimulates in vitro proliferation of both immature and IVF-GCs and inhibits FSH-stimulated E 2 synthesis in cultured GCs (Table 6). The ability of EGF to enhance both basal and hcg-stimulated P production in cultured IVF-GCs (200, 209) has not been confirmed by other

15 April, 1996 OVARIAN FOLLICULAR DEVELOPMENT IN PRIMATES 135 TABLE 6. Effects of EGF/TGFa at the follicular level Granulosa Stimulates proliferation (human follicles) (199, 200) Attenuates FSH-induced FSH-R mrna expression (rat) (201) Attenuates activin-induced FSH-R synthesis (rat) (202) Attenuates FSH-induced LH-R synthesis (rat, pig) (203, 204) Inhibits FSH-induced estradiol synthesis in immature GC (205) Inhibits FSH-induced aromatase expression in IVF-GC (206) Inhibits inhibin secretion (bovine) (207) Theca interna Inhibits camp-induced production of 17a-0H P and DHEA, and camp-stimulated P a activity (150) Inhibits LH-induced androgen synthesis (rat) (208) Dedifferentiates theca cells into stroma cells (rat) (208) When available, the reported effects come from experiments performed in primates; when not available in primates, the reported effects come from other mammal species and are in italics studies (150, 210). The reason for this is probably that the authors used different culture conditions. Both EGF (52) and TGFa (51), as well as EGF-R, have been immunohistochemically detected in human follicular tissues (Table 7). Their immunostaining increases as follicular size increases. However, these findings are not in keeping with the observed decrease in EGF and TGFa levels seen in FF as follicles enlarge ( ). Hence, intrafollicular concentrations of EGF decrease approximately 3-fold when follicles grow from 1 to 6 mm (212), and those of TGFa decrease more than 10-fold when follicles grow from 2 to 10 mm (214). In 6-mm follicles, EGF concentration is 2-fold higher than in serum (211, 212) whereas it is half the serum value in preovulatory FF (211). Taken together, available data support the view that in humans, EGF and TGFa are negative modulators of FSHinduced synthesis of E 2 and positive modulators of GC proliferation. Their production by human follicles seems to decrease as follicular diameter enlarges. 3. Basic fibroblast growth factor (bfgf). bfgf is a peptide that has been isolated from a wide range of tissues. It is known to be mitogenic for GCs in a variety of species, including humans (199), to play an inhibitory action on GC differentiation (150, 215, 216) and TIC steroidogenesis (150), and to be a potent angiogenic factor (217). In humans, mrnas encoding for bfgf and its receptor have been observed in the fetal ovary (218) and in IVF-GCs (219, 220). The RT-PCR technique used to detect these mrnas in the latter cells indicates that the message is very low. To date, the possible role of bfgf in human follicular function has received little attention. 4. TGF/3. TGF/3 is a homodimeric polypeptide with at least three isoforms (TGF/31, TGF/32, and TGFj33) that have been identified in mammalian cells (for review see Ref. 198). There are few available data concerning the action of TGF/3 on follicular tissues in primates. TGF/3 inhibits forskolininduced P450 17a activity, and 17a-OHP and dehydroepiandrosterone production by cultured TICs from human preovulatory follicles, but enhances forskolin-stimulated 3/3HSD activity by these cells (150). At the level of GCs, TGF/3 alters neither aromatase activity nor P production (150). In addition, TGF/3 enhances inhibin production by rat GCs (221), as well as binding of FSH to GCs (222). In humans, TGF/31 and TGF/32 have been immunohistochemically detected in both GCs and TICs of growing follicles from the early antral stage (Table 8). The lack of detectable TGFj3 in the FF, as shown in the porcine (225), suggests that TGF/3 actions may be only local. Hence, available data in humans support the view that TGF/3 production is predominantly associated with the largest follicles and that TGF/3 may favor production of P by TICs. 5. Inhibin, activin, and follistatin (FSP). Inhibin is a glycoprotein that has been isolated from FF as a heterodimer consisting of a common a-subunit combined with one of two /3-subunits, /3A (inhibin-a) or /3B (inhibin-b). The /3A- and jsb- subunits give rise to the three forms of activins: activin (/3A/3B), activin A (0A/3A), and activin B (/3Bj3B). FSP is a monomeric glycosylated protein that is the major highaffinity-binding protein for both inhibin and activin present in human FF (for review see Ref. 226). Among the most remarkable effects of inhibin and activin, are the opposite effects that these peptides exert on certain follicular functions (Table 9). On one hand, inhibin stimulates and activin inhibits androgen production by TICs. On the other hand, inhibin may be a potent inhibitor of ovarian cell proliferation in mice (236), while activin has been shown to stimulate in vitro proliferation of immature rat GCs (237) and of human IVF-GCs (229). However, the possible role of activin in regulating aromatase activity in primate GCs remains unclear because activin stimulates FSH-responsive aromatase activity in GCs from marmoset TABLE 7. Intrafollicular levels of EGF, TGFa, and EGF-R during folliculogenesis in primates Primary Secondary Preantral (class 1) Small antral (classes 2-4) Selectable (classes 5 and 6) Preovulatory (classes 6-8 ) Preovulatory (class 8 6 ) EGF TGFa EGF-R GC TIC FF GC TIC FF GC TIC The staining or the level in FF was determined as positive (+ to + ++). Selected follicle before the midcycle gonadotropin surge. 6 Preovulatory follicle after the spontaneous midcycle gonadotropin surge or after an hyperstimulation treatment. FF, Follicular fluid; 0, oocyte. [Data from Refs , and ]

16 136 GOUGEON Vol. 17, No. 2 TABLE 8. Follicular levels of TGF/3 during the primate folliculogenesis Follicle type Small antral (classes 2-4) Selectable (classes 5 and 6) Preovulatory (classes 6-8") Preovulatory (class 8 6 ) GC TGF/31 TIC -/ GC TGF/32 TIC TGF-/31 and /32 mrna GC The staining was determined as positive (+ to ) or negative a Selected follicle before the midcycle gonadotropin surge. 6 Preovulatory follicle after the spontaneous midcycle gonadotropin surge or after an hyperstimulation treatment. [Data from Refs. 223 and 224.] monkeys and inhibits aromatase activity in human IVF- GCs (231). Activin has been shown to play an important role in regulating the functional integrity and development of GCs in the ovarian follicle. Activin increases the number of FSH-Rs in rat GCs (202, 232) and induces reorganization of follicular structures from monolayer culture of immature GCs in both rats and humans (237). Receptors for activin as well as their mrnas have been localized in oocytes, GCs, and TICs of the rat (238, 239). However, although activin exhibits many effects on human follicular cells, mrnas for type II activin receptors (Act RII and Act RUB) have not been detected in any cell layer in 3- to 7-mm human follicles (240). The FSP effects on P production by preovulatory GCs are indirect. FSP inhibits the suppressive effects of activin (230), since it binds with high affinity to activin and with lower affinity to inhibin (241, 242). In primates, activin (234), inhibin, inhibin subunits, and FSP (Table 10), as well as their respective mrnas (240,245,246,250, 251), have been predominantly detected in GCs. However, recent findings indicate that inhibin subunits, FSP, and their respective mrnas may not be exclusively synthesized by GCs. While a-subunit mrna was detected in TICs from 1- to 8-mm human follicles, there was no immunoreactivity for the protein. Likewise, /3A-subunit and FSP immunoreactivity were localized in TICs from 1- to 8-mm follicles with no corresponding detection of mrna signals (240, 250), whereas /3A-subunit mrna was detected in TICs from dominant follicles, but the protein was not found in these cells (250). These discrepancies may be explained by several mechanisms (240) but need to be confirmed by further studies. The kinetics of production of inhibin subunits in primate GCs during folliculogenesis indicate that the activin gradient of the developing follicle declines as its inhibin and FSP gradient increases. Although the in vivo roles of these peptides are still incompletely understood, these changes could have implications for their putative regulatory functions during folliculogenesis Cytokines. Cytokines, which are primarily produced by white blood cells, constitute a group of proteins that modulate a variety of cellular functions (for review see Refs ). In humans, various ovarian effects of cytokines on cultured IVF-GC functions have been reported (Table 11). Taken together, available data indicate that cytokines may be inhibitors of GC steroidogenesis just before ovulation. Some cytokines could alter gonadotropin-stimulated production of steroids via specific receptors because receptors for interleukin-1 (IL-1) and their mrnas have been detected in IVF-GCs (263, 264). Tumor necrosis factor-a (TNFa) (255, 265), interferon-a and -7 (266), IL-1, IL-2 (260,267,268), and IL-6 (269,270) have been identified in FF from women undergoing IVF&ET. Levels of interferon-a and -7 in FF do not correlate with serum values whereas those of both IL-1 and IL-2 are half the serum value, and those of IL-6 are 3- to 30-fold higher than levels detected in serum. Messenger RNAs for IL-1/3 are absent from whole ovarian material from days 4 or 12 of an unstimulated menstrual cycle, but IL-la and IL-1/3 mrnas have been detected in preovulatory follicular aspirates obtained at the end of an IVF cycle (263, 264). Whether IL-1/3 and IL-la mrnas are produced by white blood cells, or by GCs, or both, remains unclear. White blood cells are present in large numbers in the human follicle wall at the time of ovulation, and IVF-GCs contain 5-15% of macrophages and monocytes (271, 272). This is because soon before ovulation, after the basal lamina is disrupted, thecal capillaries rapidly infiltrate the ovarian granulosa layer and come into direct contact with GCs. In addition, a protein factor that is chemotactic for neutrophils was found in human FF (273) and may facilitate infiltration of granulosa layer by these cells. Yet, when the population of monocytes is reduced by using various processes (260, 263), the "pure" GC population still contains IL-lj3. Since peripheral monocytes obtained at the time of oocyte retrieval are negative for IL-1/3 gene expression (263) and because IVF-GCs, not macrophages, account for the majority of the immunocytochemical staining for IL-1/3 in follicular aspirates (264), it has been suggested that IL-1/3 may be derived from GCs. Similarly, a recent study led to similar conclusions concerning IL-6 (269). In the human ovary, TNFa has been detected by immunohistochemistry in healthy antral follicles at the level of the antral layer of GCs (265). Because FSH has been shown to increase release of TNFa by cultured IVF-GCs possibly contaminated by macrophages, the effects of FSH on release of TNFa has been tested directly on macrophages (258). No effect of FSH was observed on these latter cells, but it is possible that FSH indirectly induces the release of TNFa. Thus, some evidence supports the view that GCs may synthesize IL-1, IL-6, and TNFa. In conclusion, although cytokines generally inhibit gonadotropin-induced steroidogenesis in cultured IVF-GCs, the physiological implications of their action remain unclear. Espey (274) has noted that ovulation involves mechanisms common to an inflammatory reaction, leading to the hypothesis that cytokines may also have a role in ovulation and luteinization rather than in cellular growth and differentiation during folliculogenesis. This hypothesis is strengthened by the observation that most cytokines appear in human follicles only a short time before ovu-

17 April, 1996 OVARIAN FOLLICULAR DEVELOPMENT IN PRIMATES 137 TABLE 9. Reported effects of inhibin, activin, and FSP at the follicular level Inhibin Granulosa Slightly inhibits FSH-induced aromatase (marmoset antral follicle) (227) Theca interna Strongly enhances IGF-I- and LH-induced androgen production (human antral follicle) (228) Inhibits activin-induced suppression of theca cell androgen synthesis (human antral follicle) (149, 227) Granulosa Theca interna Granulosa Activin Stimulates proliferation in cultured IVF-GC (229) Enhances FSH-induced aromatase expression (marmoset antral follicle) and basal and LH-induced aromatase expression (marmoset preovulatory follicle) (227) Inhibits LH-induced progesterone production (marmoset and human preovulatory follicles) (227, 230) Decreases, in a dose-related fashion, P production and aromatase activity in cultured IVF-GC (231) Increases the number offsh-r (rat) (202, 232) Stimulates FSH-stimulated LH-R synthesis (rat) (233, 234) Strongly inhibits IGF-I- and LH-induced androgen production (149) Increases basal progestin production (human preovulatory follicle) (235) Mainly inhibits activin actions (human preovulatory follicle) (235) When available, the reported effects come from experiments performed in primates; when not available in primates, the reported effects come from other mammal species and are in italics. TABLE 10. Follicular levels of inhibin, inhibin subunits, and FSP during primate folliculogenesis Inhibin Follicle type GC FF a-subunit GC Small antral (classes 2-4) Selectable (classes 5 and 6) Preovulatory (classes 6-8 ) Preovulatory (class 8 6 ) FSP /3A-subunit GC pb-subunit GC The staining or the level in FF was determined as positive (+ to + ++) or negative (-). a Selected follicle before the midcycle gonadotropin surge. 6 Preovulatory follicle after the spontaneous midcycle gonadotropin surge or after an hyperstimulation treatment. FF, Follicular fluid. [Data from Refs ] TABLE 11. Effects of cytokines on human IVF-GC TNF-a Simulates proliferation (255, 256) In synergy with IFNy or IL-1, inhibits hcg-induced P production (257) In synergy with IFNy, reduces FSH-induced E 2 production (257) Simulates PGF2 a production (255, 258) Interleukins IL-1 inhibits FSH- and hcg-stimulated E 2 production (259, 260) IL-2 inhibits hcg-stimulated P production (261) Interferons Reduce the levels of E 2 and P in FF (IVF follicle) (262) IFN-a, -/3, and -y reduce hcg-stimulated P secretion and FSHstimulated E 2 (257, 261) When administered in vivo, IFNa reduces FF and serum levels of E 2> and IFNy reduces FF and serum levels of P, in the late follicular phase without affecting gonadotropin levels (262) lation and by studies showing that cytokines display stimulating effects on various ovulation-associated phenomena (255, 258, 264, 274). VI. Atresia In mammals, it has long been recognized that the majority (>99%) of follicles present at birth become atretic with only less than 1% achieving ovulation. Thus, the "normal" fate for each individual follicle is to disappear by FSP FF entering atresia, a process that may be considered normal, allowing the ovary to produce cyclically the ovulatory quota. A. Morphological characterization of atresia In histological studies, morphological criteria have been used to assess the health status of individual follicles (i.e. whether actively growing or degenerating). Oocyte involution, irregular shape of the follicle and/or of the oocyte and, above all, nuclear pyknosis in granulosa layers (Fig. 8) are criteria most commonly used to describe early atresia (275). It is now accepted that GC pyknosis is an apoptotic process ( ). As a result of endonuclease activity, genomic DNA is cleaved into nucleosomal oligomers of about 180-bp nucleosomal units before apoptosis becomes morphologically evident (for review see Ref. 278). The first morphological evidence of apoptosis is condensation of nuclear chromatin into crescentic caps at the periphery of the nucleus, an effect associated with nucleolar disintegration and an overall reduction in nuclear size. At the same time, cytoplasmic condensation results in a reduction of total cell volume and a related increase in cell density. Atretic human follicles larger than 1 mm have been arranged into four successive stages of degeneration (Fig. 8). Whereas in follicles less than 1 mm, atresia is characterized by a fast and early oocyte degeneration as well as hypertrophy of TICs (Fig. 9), the oocyte persists long after GCs have

18 138 GOUGEON Vol. 17, No. 2 FIG. 8. Atresia in human antral follicles larger than 1 mm. A, Stage A: up to 20% of the GC of the earliest atretic follicles exhibit pyknoses and fragmentation. Theca and granulosa layers are not markedly altered, and the follicular shape remains regular with a normal basal lamina, X30, X320. B, Stage B is a further step of atresia, pyknotic GCs are numerous, and 10-50% of the granulosa wall has disappeared. The TI is not altered, and the follicular shape remains regular, X30, X320. C, Stage C: the follicle is shrunken, and the cumulus oophorus persists despite the disappearance of the GCs. TI has changed into an interstitial gland, and basal lamina is thickened with a thin strip of fibroblasts in its inner part, X30, X320. D, Stage D corresponds to the ultimate stage of degeneration; the follicle is more or less collapsed, and granulosa, theca-interstitial gland, and cumulus oophorus have disappeared. The basal lamina is very thick and the antrum is invaded by fibroblasts. When not degenerated, the oocyte is floating free in the antral cavity, X30, X320. [Reproduced with permission from A. Gougeon, The Ovary, pp (5).] disappeared in follicles larger than 1 mm and displays cyclic meiosis-like alteration (including germinal vesicle breakdown, alignment of chromosomes in metaphase, and expulsion of a polar body) that rises sharply during the periovulatory period (279). B. Follicular atresia in the human ovary during the menstrual cycle It is generally assumed that adult follicular atresia is composed of an underlying tonic component that mainly affects FIG. 9. A, Atretic preantral follicle (X280); the oocyte disappears before the TI hypertrophies. B, Atretic early antral follicle (x480). The oocyte has disappeared, and the incidence of pyknotic GCs is low but TICs are hypertrophied. [A reproduced with permission from A. Gougeon, The Ovary, pp (5).] the population of RF, reflecting a life-long process on which cyclic losses are superimposed (See Section II). In primates, atresia affects early growing follicles only slightly (10). The mean rate of atresia is moderate and does not exhibit substantial cyclic changes for the smallest growing follicles; it is similar for preantral (class 1) and early antral follicles (class 2), 30% and 32%, respectively, and for 0.5- to 0.9-mm (class 3) and 1- to 2-mm follicles (class 4), 15% and 16%, respectively. In contrast, atresia of the largest follicles varies significantly throughout the cycle, being inversely related to circulating FSH levels (105, 110). The highest incidence of atresia is observed during the midluteal phase for both the 2- to 5-mm (class 5) and 6- to 10-mm (class 6) follicle types, approximately 73% and 100%, respectively (61, 105). C. Alterations of follicular metabolism during atresia It is generally assumed that atretic follicles possess an intrafollicular androgenic milieu that distinguishes them from healthy follicles (280). However, this must be considered with caution since healthy antral follicles smaller than 8 mm possess low aromatase activity (111, 112, 281) and exhibit a high FF A/E 2 ratio ( ), grossly similar to that

19 April, 1996 OVARIAN FOLLICULAR DEVELOPMENT IN PRIMATES 139 observed in early atretic follicles of similar sizes (115). On the contrary, healthy follicles larger than 8 mm differ strongly from atretic follicles of similar size in that, for most of them, their FF A/E 2 ratio is between 0.1 and 1.1, and they possess aromatase (112), whereas the A/E 2 ratio in atretic follicles is between 1.4 and more than 40 (115), and aromatase activity is lacking or poorly expressed (112). In addition, binding of FSH to GCs of atretic follicles is patchy and reduced when compared with healthy follicles (282); therefore, as large human follicles degenerate, their GCs appear to lose their capacity to respond to FSH and, rather than being a cause, the absence of aromatase appears to be a consequence of this altered follicular responsiveness to FSH. Other alterations in follicular metabolism have also been observed in human atretic follicles. They include the appearance of lipid droplets, 3/3-HSD, glucose-6-phosphate dehydrogenase, acid phosphatase, and aminopeptidase, as well as a profound decrease in the levels of lactate deny drogenase in FF of human atretic follicles when compared with healthy follicles (283, 284). D. Causes ofatresia 1. Mechanisms leading to activation of the apoptotic cascade. The generation of oxidative free radicals (oxidative stress) is a consequence of normal cellular respiratory metabolism and reduction-oxidation (redox) reactions (for review see Ref. 285). Exposure of hepatocytes to a redox-active compound that induces oxidative cell injury activates a Ca 2+ -sensitive endonuclease leading to DNA cleavage preceding loss of cell viability (286). At the ovarian level, it has been shown, first, that GCs collected from ovaries of gonadotropin-primed rats contain high levels of a Ca 2+ /Mg -sensitive endonuclease that can be activated by increased nuclear concentrations of Ca 2+ and Mg 2+ (287), and second, that oxidative stress induces GC apoptosis (288). In addition, DNA damage, such as that initiated by oxidative free radicals, may be a primary stimulus for increased p53 (the product of the p53 tumor suppressor gene) expression in GCs (289). Since p53 can increase the transcriptional activity of death-inducer genes, and immunohistochemical studies show that p53 is essentially restricted to nuclei of degenerating GCs of atretic antral follicles, it has been suggested that p53 has the potential to amplify the negative effect of oxidative free radicals on GC viability (Ref. 289 and references therein). Taken together, these findings indicate that oxidative stress may be a key element in inducing GC apoptosis via alteration of the cellular ionic environment, which activates a Ca 2+ / Mg 2+ -sensitive endonuclease resulting in activation of the apoptotic cascade. This, however, poses the question of which factor(s) is (are) susceptible of activating /repressing stress response factors to prevent/induce follicular atresia. 2. Alteration of the tropic hormonal signal. Gonadotropins have been reported to play a critical role in preventing apoptosis in GCs of rat antral follicles in vivo and in vitro (288, 290). Consequently, insufficient or absent tropic hormonal signal may be pivotal in inducing atresia. This may occur when circulating FSH is absent or too low. Many sources indicate that hypophysectomy leads to the general atresia of all growing follicles from the antral stage onward (for review see Ref. 275), but this is not merely a physiological situation. During the menstrual cycle, circulating levels of FSH exhibit important variations: FSH is highest during the first half of the follicular phase and lowest during the midluteal phase, and during this last phase selectable follicles exhibit their highest rate of atresia. Recently, Tilly et al. ( ) provided new findings on mechanisms by which insufficient FSH support can induce atresia. By inducing synchronous apoptosis in a homogeneous population of antral follicles by culture in serum-free conditions without tropic support, these authors have shown that the ability of PMSG to inhibit apoptosis was mimicked by inhibitors of oxidative free radical formation and action, including superoxide dismutase, catalase, N-acetyl-L-cysteine, and ascorbic acid (288). These results confirm the role of oxidative stress in inducing atresia, a process that may be prevented by action of oxidative stress response factors. Among them, the product of the bcl-2 gene acts as an antioxidant or free radical scavenger (Ref. 291 and references therein). It has been hypothesized that the underlying events involved in triggering GC apoptosis and atresia would involve members of bcl-2 gene family (bcl-2, bax, bd-x ]ong and frc/-x short genes) (Ref. 291 and references therein). Briefly, the fate of a GC during development may be decided by the ratio or balance of death repressors (bcl-2 and bcl-x ]ong ) to death inducers (bax and bcl-x short ), the shift in the bcl-2 to bax ratio being primarily the result of a marked change in bax expression with essentially no change in the level of bcl-2. Apoptosis induced in follicles incubated in vitro without tropic hormones is associated with a marked increase in the amount of bax mrna; conversely, by reducing the amount of both bax and bcl-x short mrnas present in GCs, FSH has the ability to prevent apoptosis (291). Consequently, it is possible that changes in bax expression provide a primary signal that determines whether follicles are to survive (low bax) or undergo atresia (high bax). Thus, it may be that members of the bcl-2 gene family function as part of a cohort of oxidative stress response factors that must be either induced or repressed to maintain cell survival. In the absence of FSH, steroids, and growth factors (291), the resultant increase in oxidative free radicals may activate Ca 2+ /Mg 2+ -sensitive endonuclease that, ultimately, leads to cell death. Moreover, the Ca 2+ /Mg 2+ -sensitive endonuclease, which is inactive in GCs that are not dying (287), seems to be present only in GCs that are sufficiently differentiated, i.e. that are responsive to FSH. In rats, immature GCs do not express the Ca/Mg 2+ -sensitive endonuclease, but treatment with PMSG results in a rapid increase in its level in parallel with GC differentiation (287). This observation is in agreement with the view that GC apoptosis is only initiated in follicles that have undergone transition to the antral stage of development (Ref. 288 and references therein) and with the observation that in humans the atretic pathway strongly differs between follicles less than approximately 1 mm (Fig. 9) and follicles larger than 1 mm, where pyknotic GCs are predominant (Fig. 8). Consequently, it is possible that FSH, in addition to inducing differentiation of the GCs, concomitantly increases expression of both the Ca 2+ /Mg 2+ -sensitive endonuclease and oxidative stress response factors as a means to

20 140 GOUGEON Vol. 17, No. 2 protect the maturing GCs from the damaging effects of metabolically derived reactive oxygen species. 3. Other possible causes ofatresia. In addition to an insufficient tropic support, numerous mechanisms or chemical messengers have been reported to regulate GC death in various model systems. Most of them activate endonuclease activity or disrupt prevailing ionic gradients. The basement membrane must undergo continuous remodeling during folliculogenesis in order for the follicle to be able to expand in size. It is likely that the remodeling process is under hormonal control. Inappropriate hormonal stimuli or specific hormonal signals of atresia may alter the nature of the basement membrane remodeling process, resulting in breaches of the permeability barrier and either exposure of GCs to immune attacks (292) or disruption of the prevailing ionic gradients leading to activation of resident endonuclease (4). It has also been suggested that atresia may be induced by chemical messengers such as GnRH and angiotensin II (Ang II). GnRH has been reported to be an atretogenic factor in the rat ovary, because a GnRH agonist increases apoptotic DNA fragmentation (293) and inhibits GC mitosis (294). Furthermore, when cultured GCs are exposed to LHRH, there is a rapid and transient increase in intracellular Ca 2+ (295). In hamsters, Ang II treatment of follicles cultured in vitro reduces the E 2 /androgen ratio and induces morphological changes in the TI that are consistent with atresia (296), and in cattle, it has been suggested that an inappropriate local production of prorenin, the precursor of Ang II, may play a role in follicular atresia leading to the inhibition of GC aromatase activity (297). In humans, however, GnRH, GnRH receptors, and their mrna have been detected in/on GCs from preovulatory follicles only ( ). Similarly, GCs from healthy follicles in normal women immunostain negatively for renin and Ang II until the midcycle gonadotropin surge (301). Taken together, these findings suggest that neither GnRH nor Ang II can act as atretogenic factors in growing nonovulatory human follicles. In rats, it has been observed that androgens enhance follicular atresia in vivo (for review see Ref. 275) and, more recently, that androgens activate endonuclease activity and enhance GC apoptosis (277). However, these experiments must be considered with caution since they have been performed in hypophysectomized, diethylstilbestrol-primed, immature rats, a unique situation in which FSH is absent. In humans, it has been shown that in follicles where A production exceeds GC-aromatizing ability, A is metabolized to dihydrotestosterone (DHT) and 5a-androstanedione, irrespective of the amount of FSH present (110). It has been speculated that DHT and 5a-androstanedione induce atresia by blocking the FSH-stimulated aromatase activity by cultured GCs (281). However, this assumption must be considered cautiously since it has been shown that DHT enhances in vitro FSH-induced aromatase activity in GCs from small antral marmoset selectable follicles (302). Taken together, available data indicate that there is no definitive evidence concerning a possible role of androgens in inducing atresia during in vivo primate folliculogenesis In rats, IGFBP-4 has been hypothesized to be involved in the induction of atresia of selectable follicles not selected to become dominant (303). Taking into account the putative role of IGFs in regulating the differentiation and proliferation of GCs and TICs, it may be expected that locally produced IGFBPs can sequester IGFs in the microenvironment and. direct nonselected follicles toward atresia. In humans, IGFBP-2 and -4 have been reported to be higher in presumptive atretic follicles (high A/E 2 ratio) than in large (>8 mm) healthy nondominant follicles (low A/E 2 ratio) (170, 304). However, these findings do not necessarily signify that high > intrafollicular concentrations of IGFBP-2 and -4 are necessary and sufficient for follicular atresia. High levels of IGFBP-2 ' and -4 mrnas are present in healthy 0.5- to 1-mm follicles (162), and IGFBP profiles do not differ between normal follicles smaller than 8 mm with a high A / E 2 ratio, PCOS, and atretic follicles of similar sizes ( , 304). We hypothesize that high levels of IGFBP-2 and -4 in large atretic follicles may be a consequence, rather than a cause, of the process by. which they have not been selected to ovulate. In conclusion, although a coherent picture on the mechanisms that may lead to activation of the apoptotic cascade is emerging, little is known of the biochemical and molecular pathways used by various factors to control the initiation of «apoptosis. It is possible, however, that they act either by repressing expression of oxidative stress response factors or by favoring expression of death-inducer genes. The early hypothesis (305), that only follicles reaching a critical stage of their development coincidentally with the appropriate changes in serum gonadotropins and / or other factors will be spared from atresia and reach ovulation, remains fully justified in that FSH and certain growth factors have been dem- onstrated to act indirectly as antioxidative molecules (288). VII. Synthetic Views on the Control of Follicular Development in Primates In this section, we will first examine the changes associated with entrance of secondary follicles into the preantral stage and the possible physiological implications in the subsequent steps of follicular growth. Second, we will propose answers to some intriguing questions in relation to folliculogenesis, using data from morphological, immunohistochemical, and biochemical studies that will be integrated and analyzed in the light of in vitro studies. A. Entry of follicles into the preantral class during the early luteal phase: physiological implications Secondary follicles become preantral when at least one TIC appears to be epithelioid. In humans, follicles enter the preantral stage throughout the menstrual cycle, but it has been observed that it is during the early luteal phase that the number of preantral follicles is highest and their mean size the smallest, reflecting a recent increased entry into this class. In addition, their thecal vascularization is significantly more developed than at any other time during the cycle. 1. Which factors are involved in the increased vascularization of preantral follicles? Angiogenesis is a complex developmental process that undoubtedly results from multiple interactions t '

21 April, 1996 OVARIAN FOLLICULAR DEVELOPMENT IN PRIMATES 141 (for review see Ref. 217). A role for LH has been suggested since in humans, hcg increases the influx of blood within follicles (67). The angiogenic effect of LH may be exerted via stimulation of Ang II production by preovulatory GCs (306) that, in turn, may stimulate bfgf secretion (307), one of the most potent angiogenic factors (217) to be detected in the human ovary ( ). In addition, FF from follicles sampled during the follicular phase in unstimulated menstrual cycles possess a mitogenic potential for bovine aortic endothelial cells (308), and IVF-GCs can synthesize an endotheliotropic factor in vitro (309). Messenger RNA coding for vascular endothelial growth factor has been identified only in GCs of antral and preovulatory monkey follicles (310). This protein induces angiogenesis under in vivo conditions and has been suggested to be stimulated by FSH (6, 217). It is also important to note that many substances involved in follicular rupture and produced by the ovulatory follicle after the gonadotropin surge are angiogenic factors as well (for review see Ref. 217). Other angiogenic factors can be produced by the preantral follicle itself. This is the case for TGFa, which stimulates endothelial cell proliferation in hamsters (311) and has been detected at the theca level and in GCs as early as the primary stage in humans (51). Available data, therefore, indicate that angiogenic factors may originate from follicles of various sizes, from the secondary to the preovulatory stage. Although the mode of action of these peptides remains unclear, they may act locally or via the general circulation to enhance thecal vascularization of preantral follicles during the periovulatory period. 2. Is there a coincidence between the periovulatory period and the differentiation oftics? Although secondary follicles can transform into preantral follicles throughout the menstrual cycle by acquisition of epithelioid cells in the theca layer (i.e. differentiation of TICs), there is a significant increase in the number of preantral follicles during the early luteal phase. LH has been considered to be the primary hormone regulating the differentiation of TICs (312). However, some observations in rats suggest that a regulatory factor distinct from LH stimulates thecal differentiation and that IGF-I, in stimulating the initial expression of LH-R during the early stages of thecal differentiation, may be this factor (313). In addition, observation of atretic preantral follicles shows that TICs hypertrophy soon after disappearance of the oocyte in humans (Fig. 9) as well as in rats where these cells develop the ultrastructural and functional features characteristic of steroidogenic tissues (314). These findings suggest that in healthy follicles a factor may exert an inhibitory influence on theca cells to prevent their differentiation. Potential candidates are EGF and TGFa, which have been detected in human oocytes and GCs, respectively (50-52), and have been postulated (208) to function as theca-organizing factors that attract theca cells around the follicle while maintaining them in an undifferentiated stage. Thus, differentiation of the TI that occurs during the early luteal phase in a higher number of preantral follicles than at any other time during the menstrual cycle may result from an increasing amount of extraovarian IGF-I reaching secondary follicles via increasing vascularization. By its ability to induce differentiation of protheca cells, IGF-I may counteract, at least partly, the inhibitory influence exerted by GCs and /or the oocyte on this differentiation. 3. Do changes represent protective mechanisms of a particular wave of follicular growth? In the mouse, it has been shown (315) that the passage of a follicle toward atresia or ovulation is determined by the time of the cycle when follicles reach stage 5a (similar to the human preantral stage). Moreover, Greenwald and Terranova (316) suggested that hamster follicles developing in synchrony with FSH and LH surges recurring every 4 days constitute the group selected to ovulate after four or five cycles of development. In both hamsters (317) and rats (318), FSH stimulates follicular production of EGF and EGF-R, and in rats, it has been shown that GC apoptosis is prevented by a tonic inhibition of endonudease activity by the paracrine/autocrine action of EGF/TGFa (290). Because TGFa and EGF are present in both TICs and GCs in preantral and early antral human follicles, respectively (Table 7), it is possible that follicles that have differentiated their TI in an optimal gonadotropic environment, e.g. during the periovulatory period, possess the ability to synthesize sufficient amounts of EGF/TGFa to be better protected against atresia than follicles that have differentiated their TI at other times during the cycle. These latter follicles may constitute the "minor" waves that normally do not lead to a preovulatory follicle; this hypothesis is in agreement with the early Richards' hypothesis (126) of an "inadequate" differentiation of the TI resulting in an imbalance detrimental to follicular development in subsequent stages of growth. From these findings, we speculate that, in women, follicles differentiating TI during the periovulatory period might be endowed with selective advantages for further normal development by periovulatory gonadotropin changes acting via increased blood vascularization. These follicles may constitute the "major" wave from which follicles will become selectable 70 days later. Nevertheless, whether all or only some members of the major wave are equally privileged (endowed with selective advantage) to become selectable is not known and requires further study. The existence of such waves could also explain why the passage of follicles from one class to the next is detected only once per cycle (see Section III.C). Inasmuch as they are less susceptible to atresia than follicles from minor waves, follicles from major waves will be found in larger numbers throughout the successive cycles, which explains why their passage through subsequent classes is discernible (Fig. 10). B. Basal follicular growth In the human ovary, the number of follicles with a small antral cavity peaks during the late follicular phase. Because the primary stimulus for antral transition is believed to be FSH (4), and because FSH alone is sufficient to induce antrum formation in human preantral follicles cultured in vitro (108), this increased entry of follicles into the early antral class may be related to the high levels of FSH observed during the first half of the follicular phase. In addition, it has been shown in rats that antrum formation requires the presence of the oocyte and that the action of FSH is enhanced by activin A (237). However, the precise mechanisms by which activin and the oocyte may be involved in antral transition remain unknown.

22 142 GOUGEON Vol. 17, No. 2 - EL ML LL EF MF LF EL ML LL EF MF LF EL ML LL EF MF LF Stages of the cycle FIG. 10. Description of the progression of the "major wave" throughout successive cycles, from the entry of follicles into the preantral (class 1) to the preovulatory stage (class 8). Follicles that became preantral during the early luteal phase of every menstrual cycle have been endowed with selective advantages for further development. As growth is progressing, the initial number of follicles belonging to the major wave {bold line) decreases by atresia, but in lesser proportion than do follicles from "minor waves" {dotted thin line). Consequently, when follicles belonging to major waves and of a given stage of their development {e.g. class 1) enter the subsequent class {e.g. class 2 during the late follicular phase) they are in larger numbers than at any other time during the cycle, thereby explaining why this passage is morphologically discernible. 1. How do follicles develop during basal growth despite their low responsiveness to FSH? We have seen above (Section IV) that follicles undergoing basal growth are relatively insensitive to cyclic gonadotropin changes in terms of growth rate and steroidogenesis. In rat GCs, low and constant doses of FSH increase FSH-R mrna levels, whereas a single high dose of either FSH or LH/hCG decreases FSH binding and FSH mrna levels (319). Because it has been observed that FSH fails to induce desensitization of adenylate cyclase in GCs of preantral rat follicles (130), it can be assumed that this property of immature GCs may be required to permit a continuous nondisruptive pattern of follicular growth when small follicles are repeatedly exposed to gonadotropin surges. To explain why follicles can grow during basal growth despite this low responsiveness to FSH, a role for oncogenes and growth factors can be suggested. The c-myc mrna is abundantly expressed in GCs from 1- to 2-mm as compared with 3- to 11-mm porcine follicles (320), and only GCs from preantral follicles immunostain positively for the myc protein in humans (55). Evidence suggests that c-myc oncogene expression is linked to growth control in the prereplicative phase and, as a result, the intriguing possibility of c-myc oncogene participation in the "autonomous" growth of the immature follicles during the early stage of folliculogenesis has been proposed (320). In addition, EGF/TGFa and IGFs have well known effects on human GC proliferation (Tables 5 and 7). Because EGF/TGFa and their receptor, as well as IGF-R, have been detected in preantral and small antral human follicles, it is likely that these factors may play a positive role in sustaining growth of small follicles. 2. Why is steroidogenic activity low in follicles during basal growth? The steroidogenic activity of follicles during basal growth is very low. A major cause for this appears to be the lack of steroidogenic enzymes in TICs (P450 scc and P450 17a) and GCs (P450 arom ). When the physiological effects of EGF/ TGFa (Table 6) are integrated with their developmentally regulated production, it appears that these peptides have the ability to block follicular steroidogenesis during basal growth, by inhibiting the FSH-induced aromatase activity in GCs and the LH-induced androgen production in TICs. The underlying mechanism of inhibition by EGF of the FSHinduced aromatization is not fully understood, but it has been reported that EGF reduces camp formation by reducing adenylate cyclase activity (321). Similarly, in TICs, EGF inhibits camp-stimulated P450 17a activity (150). TNFa, immunohistochemically localized in small healthy antral follicles of the human ovary (265), may also be involved in the inhibition of steroidogenesis, because of an existing dosedependent inhibitory effect of TNFa on basal or LH-stimulated A production by TICs from immature rats (322). In addition, it has been suggested in rats (323) that TNFa may be an important modulator of the gonadotropin-induced transition of GCs from a proliferative to a differentiative state. Although the cellular mechanism of this regulation is yet to be determined, it is possible that the mitogenic influence of this cytokine, and maybe of EGF/TGFa, is dominant over the differentiated function of FSH and, as such, promotes the progression of the GCs from the G1-G0 to S phase of the cell cycle and results in the "switching off" of FSHinduced differentiation. C. Selectable follicles Selectable 2- to 5-mm follicles are present throughout the cycle. In contrast to smaller follicles they are sensitive to cyclic changes in FSH since both the percentage of healthy follicles and mitotic index of their GCs are low when FSH is low and are high when FSH is high. In addition, the hmginduced GC proliferation is highest when women are treated

23 April, 1996 OVARIAN FOLLICULAR DEVELOPMENT IN PRIMATES 143 during the late luteal phase (114). However, although their TICs synthesize significant amounts of aromatizable androgens, aromatase activity of their GCs and subsequent E 2 production (Table 3) remain very low. 2. How do these follicles respond to changing levels of FSH by altering GC proliferation? Both activin and EGF, present in primate selectable follicles, stimulate IVF-GCs (200, 229) and / or immature GCs (199) to proliferate. The highest levels of inhibin /3 B -subunits have been observed in selectable monkey follicles (245). Therefore, because treatment of GCs with FSH in rats enhances the number of activin-r (324), and in vivo treatment with FSH increases the ovarian EGF-R content approximately 2-fold (318), it is possible that activin, acting alone, and EGF, acting in synergy with FSH, stimulate proliferation of GCs in human selectable follicles when FSH increases. Follicular vascularization may also be involved in this process. In rabbits, it has been observed that the number of thecal capillaries increases linearly with increasing follicular size (325), thereby increasing the blood flow rate and consequently the amount of both FSH and IGF-I reaching the follicle. Consequently, it appears likely that in human selectable follicles, activin, EGF, and IGF-I, acting alone or synergistically (326), have the potential to increase the FSHinduced GC proliferation when FSH levels increase, especially during the late luteal phase. 2. Why does production ofe 2 remain low in selectable follicles? The mean concentration of A in FF from selectable follicles is of 638 ± 113 ng/ml whereas that of E 2 is 15 ± 5 ng/ml (Table 3). TICs of these follicles do not stain positively for P450 scc and P450 17cv, suggesting that these enzymes are present in low quantities. However, these cells possess IGF-R (Table 5) and more LH-R than those of follicles during basal growth (123, 124). Consequently, extraovarian IGF-I and LH, which have positive effects on androgen production by TICs (89, 151), may, via increasing theca blood vascularization, stimulate biosynthesis of A and counteract the negative effects of both activin (149) and EGF (150,208) on androgen production by TICs. Although both FSH and IGFs stimulate in vitro production of E 2 by GCs independently of cell maturity (Table 4), and despite the presence of IGF-R on GCs (Table 5), these follicles synthesize very low amounts of E 2. Consequently, factor(s) inhibiting the aromatization of androgens by GCs may be operating inside the follicle. This is the case for EGF and TGFa, which are present in selectable follicles (Table 7). In addition, IGFBPs, and especially IGFBP-2 and -4 (49), may inhibit the positive action of intrafollicular IGF-II on E 2 production (148). It can be suggested, therefore, that EGF/TGFa and IGFBPs have additive effects that inhibit FSH-induced aromatase activity at the GC level. Alternatively, it is possible that the "FSH threshold" is lower for inducing GC proliferation than for inducing cytodifferentiation. Continuous in vivo thymidine infusion to rats indicated that within each antral follicle the granulosa layer is composed of two distinct subpopulations: proliferating cells and nonproliferating cells undergoing differentiation (327). Thus, it is possible that because of their prevailing proliferative status, GCs in selectable follicles exhibit a very high stimulation threshold for aromatase expression, which may require either a high dose of a single factor or a combination of effects from multiple factors to activate the gene responsible for synthesis of aromatase protein. D. Selection of the follicle destined for ouulation Every mammalian species is characterized by a fixed number of follicle(s) that ovulate at each cycle (ovulatory quota). The term selection has been used by Goodman and Hodgen (75) to indicate the final adjustment of the cohort to a size equal to the species-characteristic ovulatory quota. At the beginning of the follicular phase the largest healthy follicle appears to be the selected follicle. It differs substantially from selectable follicles in the A/E 2 ratio in FF that is comprised between 0.1 and 1 (Table 3). Thus, one of the most important events that has occurred during selection is the appearance of aromatase activity in GCs. 1. What regulates changes in steroidogenic activities in the newly selected follicle? In rats, it has been observed that high concentrations of FSH inhibit IGFBP-4 production and increase IGFBP-4 protease activity that hydrolyzes IGFBP-4 into smaller fragments that do not bind the IGF ligands (328). In humans, both IGFBP-2 and -4, which were at high levels in small follicles (49), strongly decrease in preovulatory follicles (171). In addition, GCs from preovulatory follicles produce IGF-II (164) and contain mrna for IGF-II at significant levels (159). As IGFs are known to act in synergy with FSH to enhance the FSH-induced aromatase activity (Table 4), it can be proposed that in the newly selected follicle free IGF-II may stimulate aromatase activity in synergy with high circulating levels of FSH characterizing the intercycle period. In addition, although the ability of activin in enhancing FSH-induced aromatase activity in GCs is disputed (231), the observation that activin stimulates aromatase in both immature and preovulatory monkey follicles (227) suggests that this peptide may reinforce the action of IGF-II at the time of selection. At the TI level, the negative effect of activin on production of aromatizable androgens may be counteracted, at least partly, by both a higher number of LH-R on TICs of selected follicles (123) and an increasing LH pulse frequency during the early follicular phase (329). 2. How is the follicle destined to ovulate "chosen"? Are these follicles different because they are derived from a different clonal pool? It has been suggested that the selected follicle is the one that is growing most rapidly in response to the intercycle rise in FSH and has differentiating GCs: i.e. the one with the lowest FSH "threshold" (61). As demonstrated in the marmoset monkey, GC aromatase activity in preovulatory follicles is at least 10 times more sensitive to FSH than in small follicles (302). The increased responsiveness of GCs to FSH in terms of aromatase activity may be related either to the existence of different transducers to which the FSH-R can couple or to changes in activity of the adenylyl cyclase enzyme through the turning on/off of expression of specific genes (122). Interfollicular differences in ability of these specific genes to be activated or repressed by various molecules, including growth factors, may be involved in the process by

24 144 GOUGEON Vol. 17, No. 2 which one follicle, being more receptive than others to FSH, is "chosen" and begins its preovulatory maturation. The character (stochastic vs. deterministic) of these mechanisms remains unknown. That the population of follicular cells (GCs and TICs) present in a given follicle constitutes a mixture of clones that have evolved from different stem cells (330) is probably of great importance. The average number of clonal progenitors has been estimated to be five in mouse GCs (331) and corresponds closely to the numbers of observed pre-gcs surrounding the oocyte in primordial follicles. Consequently, in a follicle that has begun its growth at a given time, the population of GCs is probably different from that of another follicle that started its growth at the same time. GCs of follicles presenting the same morphological characteristics may display differences in their development potentialities and capacities for synthesizing molecules that affect their own growth and differentiation. It is therefore tempting to speculate on the existence, within a population of healthy follicles of similar size, of some follicles presenting functional differences, even if moderate, that lead to a prompter expression of genes governing the synthesis of factors or enzymes favoring growth and ultimately full differentiation. This paradigm may explain why, after several weeks of growth, a given follicle rather than another is selected for ovulation. Being capable of appreciating even minimal differences in the expression of genes controlling the synthesis of molecules involved in follicular development will probably help answer this question. E. Preovulatory maturation Between selection and the time of the plasma E 2 peak, the mean steroid concentration in preovulatory FF increases 2-fold. Interestingly, preovulatory FF before the midcycle gonadotropin surge exhibits significant levels of progestins (Table 3). Because GCs from human follicles before the midcycle gonadotropin surge lack 3/3HSD in vivo (113,121), progestins observed in FF are probably of thecal origin. This may be due to an effect of TGF/3, whose production is highest in preovulatory TICs (Table 8), and has been reported to enhance forskolin-stimulated 3/3HSD activity by cultured TICs from human preovulatory follicles (150). 1. What is the role of the inhibin-activin-fsp system during preovulatory maturation? Studies of FF levels of androgen and estrogen in relation to GC aromatase activity indicate that the capacity of the TI to generate A increases in parallel with aromatase activity in the human preovulatory follicle. It has been suggested that a positive feedback loop (GCs on TICs) promotes thecal androgen synthesis and hence estrogen synthesis in this follicle (94). The kinetics of production of inhibin subunits in monkey follicles during their growth indicates that the activin gradient of the developing follicle declines as its inhibin gradient increases (245). In addition, throughout preovulatory maturation, FF contains increasing levels (332,333) of FSP, whose production is under the control of FSH (334). FSP, which preferentially binds activin (242, 335), inhibits its biological effects (235). While the inhibitory effect of activin on production of aromatizable androgens by TIC is progressively raised, preovulatory GCs produce increasing amounts of inhibin (247, 336), which strongly enhances the LH- and IGF-stimulated androgen production by TICs (228) and suppresses the inhibitory action of activin on androgen synthesis * by these cells (149,227). Regarding the FSP action, it is tempt- / ing to postulate that this molecule may serve to inhibit or modulate the action of a substance whose synthesis cannot be blocked specifically. During preovulatory maturation, increasing amounts of inhibin are being synthesized, a phenomenon dependent upon the synthesis of a- and j3-sub- \ units. Therefore, any substance specifically blocking activin synthesis, whose inhibiting effects on TI androgen synthesis could hamper the development of the follicle destined to ovulate, will also block that of inhibin, thereby confusing the overall effects of the blocking substance. By binding more. specifically to activin than inhibin, FSP can strongly inhibit activin action without substantially affecting that of inhibin. * These observations show that, as the preovulatory follicle \, matures, it produces increasing amounts of inhibin and FSP, which have a positive effect on androgen synthesis by TICs, either directly or indirectly. The biological effects of inhibin and activin on follicle cells and their changing expression during follicle growth indicate, that they may play a key role in folliculogenesis. Studies by Matzuk et al. (236), who used inhibin-deficient mice, indicated that in the absence of inhibin, mice can develop into young adults that are not completely normal. Within a few weeks, highly invasive and hemorrhagic tumors that resemble GC tumors develop in the ovary. Whether this abnormal ovarian cell proliferation is due to lack of inhibin or to hypersecretion of FSH is not known. In addition, these animals are infertile. The reason for infertility in these inhibindeficient mice is not yet understood; the authors suggested < that it is secondary to the early development of the tumors either due to hypersecretion of FSH or to the absence of inhibin (236). Although development of normal secondary sexual organs and mature-appearing gametes have been reported, a review of the accompanying microphotographs ^ shows that the ovary from a 4-week old inhibin-deficient animal does not resemble the ovary of a normal animal of the same age in that it contains no antral follicles. Therefore, it is possible that infertility is due not only to the occurrence of tumors but also to an impaired follicular growth. 2. Why does aromatase activity increase before the midcycle gonadotropin surge? FSH is the primary stimulant of functional A GC differentiation at the onset of the preovulatory phase (337). At the level of GCs, many substances have been reported to cooperate with FSH to stimulate aromatase expression. T and DHT enhance in vitro FSH-induced aromatase activity in GCs from small antral marmoset selectable» follicles (302). They may act via androgen receptors present in human follicle cells (338, 339). In rats, E 2 stimulates aromatization of androgens in GCs (126). Although estrogen receptors have been detected in large antral follicles of primates during the follicular phase ( ), no positive effect of E 2 on FSH-induced aromatase activity has yet been reported in primate GCs. In primates, estrogens probably do not play a key role in folliculogenesis as assessed by the > presence of ovulatory follicles after treatment with exoge-

25 April, 1996 OVARIAN FOLLICULAR DEVELOPMENT IN PRIMATES 145 nous gonadotropins in patients suffering from a 17a-hydroxylase deficiency (343) and by the presence of follicular growth after treatment with recombinant FSH in LH-deficient monkeys despite production of E 2 being strongly im- >^ paired (84). During preovulatory maturation, production of IGF-II and its mrna by GCs increases; FF concentration of IGF-II increases (Table 5), while those of IGFBP-2 and -4 strongly decrease. It can therefore be assumed that IGF-II may be the most effective stimulator of GC aromatase ac- 4 tivity in synergy with FSH which is also involved in stimulating the expression of GC IGF-R (3). Taken together, these findings indicate that as the preovulatory follicle matures, it produces increasing levels of molecules having a positive effect on production of E 2 by GCs. [ 3. Why does GC proliferation decrease before the gonadotropin surge? Is this change due to GC differentiation? When the prei ovulatory follicle begins to produce high levels of E 2, the rate of cell proliferation in the membrana granulosa begins to decline (66). This change, which parallels an increase in the capacity of GCs to produce E 2, i.e. to differentiate, can be explained by the progressive change in the ratio of prolif-»> erating GCs to nonproliferating GCs undergoing differentiation. Thus, in the newly selected follicle some GCs differ- ' entiate, as compared with selectable follicles, and begin to produce E 2 while FSH, acting in synergy with factors favoring proliferation, stongly stimulates the growth of a large number of remaining proliferative GCs. The more the follicle matures, the less there are GCs that retain their capacity to proliferate. FSH, acting in synergy with cytodifferentiative ^ factors, causes more GCs to differentiate and express the FSH-induced functions resulting in a simultaneous decrease * of the GC proliferation and increase in E 2 production. At this time, adenylyl cyclase, which is highly responsive to FSH (130), produces high levels of camp, which has been shown to inhibit GC proliferation in cattle (344), possibly by acting on cell-cycle factors. 4. How does the preovulatory follicle exert its dominance upon less developed follicles? The follicle destined to ovulate has been termed the "dominant follicle" (345), a concept originating from the observation of a bilateral arrest in the development of the less mature follicles in the presence of the preovulatory, follicle. Thus, destruction of the dominant follicle during the late follicular phase is followed by a gonadotropin surge ^ occurring 2 weeks later, i.e. the time interval necessary for a selectable follicle to reach the preovulatory size, in both monkeys (75) and women (76). This delay corresponds with the situation prevailing in humans, where the mean diameter of the largest nonovulatory healthy follicle is 2.4 ± 0.7 mm * during the late follicular phase (60). In addition, it has been shown that the hmg-induced GC proliferation in selectable follicles decreases as the time of ovulation approaches (114) and that the hmg-induced follicular growth is suppressed in the presence of the preovulatory follicle (346). These data confirm that in the presence of a preovulatory follicle, the mechanisms leading a follicle to be selected are inhibited. DiZerega et al. (for review see Ref. 347) assumed that domi inance originates from the secretion of follicular substance(s) termed "follicular regulatory protein(s)," which have the ability to suppress the follicular response of less developed follicles to gonadotropins by inhibiting their gonadotropinstimulated aromatase. To date, the exact nature of follicular regulatory protein has not been elucidated. This hypothesis of a direct inhibition of less-developed follicles by the preovulatory follicle has been disputed by others who suggested an indirect dominance via increasing circulating levels of E 2 (348, 349) and subsequent fall in FSH occurring from the midfollicular phase. 5. How does the preovulatory follicle continue to thrive when FSH levels are decreasing? Although FSH appears to be essential for the final growth of the follicle destined to ovulate (349), its GCs bind less FSH than do those of early preovulatory follicles in midfollicular phase (124, 282), and from the midfollicular phase onward, the circulating levels of FSH progressively fall in response to increasing plasma E 2 levels. Despite less circulating FSH, the amount of FSH reaching the dominant follicle may be either unchanged or increased since it has been shown (69) that during the second half of the follicular phase, the preovulatory follicle is much more vascularized than are less developed follicles. Production by the preovulatory follicle of angiogenic factors, including vascular endothelial growth factor (310), bfgf (219, 220), Ang II (306), cytokines (217), or unidentified factors (309, 350), may account for the increase in thecal vascularization. Thus, a causal relationship between gonadotropin stimulation and the expression of angiogenic substances suggests an intraovarian mechanism by which the dominant follicle protects itself from the fall in plasma FSH concentrations. That is, by developing an elaborate vascularization network, the dominant follicle may preferentially increase its exposure to gonadotropins. Moreover, a change in FSH-R function, or peptides actively synthesized by the dominant follicle itself, may play a basic protective role by enhancing sensitivity of GCs to FSH. It has been demonstrated in the marmoset monkey that GC aromatase activity and P accumulation are at least 10 times more sensitive to FSH in preovulatory follicles than in small follicles (302, 349), and in rats, GCs isolated from preovulatory follicles exhibit greater FSH-responsive adenylyl cyclase than GCs isolated from small antral follicles in which steroidogenesis is low (130). It has also been suggested that the presence of LH-R on preovulatory GCs, which appear during the midfollicular phase (124, 127) under the primary control of FSH acting in synergy with E 2 (126), TGF/3 (222), and IGFs (351), could serve to protect the follicle from the fall in plasma FSH concentrations by allowing the cells to produce camp. Thus, an increase in angiogenesis and sensitivity of GCs to LH and FSH may explain the maintenance of the preovulatory follicle development during the late follicular phase when circulating levels of FSH are decreasing. F. Preovulatory growth after the midcycle gonadotropin surge After the midcycle gonadotropin surge, the preovulatory follicle switches from a E 2 -producing to a progestin-producing structure. Its production of steroids strongly increases, rising from a mean FF concentration of near 4,800 ng/ml

26 146 GOUGEON Vol. 17, No. 2* before the surge, to nearly 11,000 ng/ml after (Table 3). In addition, GC proliferation is arrested and capillaries from the theca layers invade the granulosa. 1. What controls the shift in preovulatory steroidogenesis after the gonadotropin surge? Although FSH and growth factors, such as IGFs (141, 144, 147), and FSP (235) in humans and TGF/3 (352) in rats, have been demonstrated to stimulate progestin production, LH is primarily involved in the shift of steroidogenesis observed after the midcycle gonadotropin surge. The decrease in intrafollicular E 2 and A after the midcycle gonadotropin surge could be due to an LHinduced inhibition of the thecal P450 17a/lyase complex since direct measurements on human GCs during the gonadotropin surge (353) or just before ovulation (354) have shown that aromatase activity remains high. Moreover, GCs from preovulatory follicles in marmoset monkeys express an LH-responsive aromatase system (355). It is evident that P450 17a/lyase is not completely inhibited because both 17a-OHP and E 2 remain present in FF at significant levels (115). The decreased secretion of androgen and estrogen observed after the LH surge may be, at least in part, accounted for by a decrease in mrnas for P450 17a and/or P450 arom in both bovine (356) and rat (357) GCs. However, it is important to note that the decline in aromatase mrna in the differentiated GCs is not necessarily associated with a corresponding loss of aromatase activity, the half-time of the enzyme being much longer than that of its mrna (357). Alternatively, an alteration of the 17j3HSD activity may be involved because data in rats suggest that the ovulatory surge of LH can down-regulate the expression of this enzyme (358). Concomitantly to the decrease in E 2 and A production in GCs after the gonadotropin surge, P production increases markedly (Table 3). This results from two additive effects. First, both 3/3HSD (66, 113, 121) and P450 scc (111) appear in GCs after the gonadotropin surge. Second, as a result of the breakdown of the basal lamina, the cholesterol substrate required by GCs for progestin production, and which is provided to cells in the form of lipoprotein-bound cholesterol, can now reach the follicle via the blood supply (for review see Ref. 64). 2. Which mechanism arrests mitoses in GCs after the gonadotropin surge? In addition to the possibility that after the LH surge GCs having completed the final step of their maturation become incapable of proliferating, a direct action of P to block GC mitosis has been suggested. At concentrations similar to preovulatory intrafollicular levels, P directly inhibits the proliferation of cultured IVF-GCs, an effect not mimicked by E 2, T, DHT, or dexamethasone, and which cannot be overridden by EGF (359). Aminoglutethamide, which blocks P secretion, increases the number of cultured IVF-GCs (359). The growth-inhibiting effect of medium containing high levels of P is removed by either RU 486 or charcoal extraction, but not by heat inactivation (360). After the gonadotropin surge, primate GCs possess P receptors (340, ), so it appears likely that P plays a role in inhibiting GC proliferation. VIII. Conclusions and Perspectives A simultaneous analysis of morphological and functional changes of follicles during their development has allowed the construction of a coherent picture concerning the mechanisms, sometimes redundant, that may control ovarian folliculogenesis in primates. Yet, this reviewer is fully aware that many pieces of this puzzle remain schematic and speculative for several reasons. First, only peptides either present in primate follicles or having an effect on primate follicle cells have been considered in the current review. This does not signify that othe# peptides present in large number in the mammalian ovary are not involved in regulation of folliculogenesis, and their possible role requires further studies. Second, the cellular signaling pathways in follicle cells and the hormonal control of gene expression in the ovary ar<$ highly complex and have not been integrated in our model. A better understanding of follicular growth needs to take into account these mechanisms (for review see Ref. 364). Finally, we must wonder about the exact role played by growth factors in the regulation of ovarian folliculogenesis. One of the most fascinating features of the intraovarian reg-. ulatory network controlling follicular growth is the redundance of systems that assure the proper function of the ovarys i.e. the production of one fertilizable oocyte every cycle. For both the control of follicular development and destruction of excess follicles through atresia, there is probably more than one single triggering factor. The molecules involved in keeping the follicular tissue in an undifferentiated state, such as those responsible for the acquisition of gonadotropin responsiveness by these same tissues, never act alone. In each step i there are several actors involved. As a result, a deficiency in^i any one of the players controlling one or several functions may be compensated for by the action of other substances. Nevertheless, despite many in vitro experiments arguing in favor of a major role played by intraovarian factors in the regulation of folliculogenesis, a significant doubt persists as» to the true physiological role they play in vivo. Because all these substances are ubiquitous and are involved in the functioning of numerous other tissues, any attempts at clarifying their function through their injection or neutralization through antisera are rendered very difficult, if not illusory. By using targeted deletion of gene (knock-out studies) governing production of various growth factors [inhibin a-subunit (365), activins (366), activins and activin receptor type 11^ (367), FSP (368), TGF/31 (53)], it has been possible to analyze the effect of these factors on embryonic development and ) reproductive function. However, one problem raised by transgenic animals is that of their viability. For example, although activin-j8a-deficient mice progress to term at they expected frequency, they died within 24 h of birth, eliminating the possibility of analysis activin effect on cyclic folliculogenesis (366). Another problem is linked to the action > of peptides of the inhibin family at both the pituitary and ovarian levels. Thus, while in inhibin-deficient mice, the circulating levels of FSH are markedly increased (365), they are strongly decreased in activin-deficient mice (367). Consequently, it is likely that impaired ovarian function is due A to abnormalities in circulating levels of FSH. However, it

27 \April, 1996 OVARIAN FOLLICULAE DEVELOPMENT IN PRIMATES 147 appears impossible to conclude whether absence of the growth factor at the ovarian level is responsible, at least in part, for the observed defects. It appears, therefore, that the \\se of knock-out transgenic mice to improve our knowledge of growth factor effects in ovarian function is limited, since these peptides are ubiquitous factors and are involved in the functioning of numerous other organs. Alternatively, the use of two other transgenic mice models may be of greater interest. In animals bearing a dominantnegative mutant protein, the overexpression of the mutant protein can be induced by the use of a tissue-specific pro- ^moter-enhancer element leading to an experimental deprivation of the normal protein (369). In animals bearing a human TGFa DNA under the control of metallothionein-1 ^promoter, the TGFa gene can be either acutely or chronically overexpressed (370). However, because there are differences femong species (rat, 371; and human, 51,214) with respect to the presence and function of TGFa in the ovary, this latter study raises the problem of specificity of action of growth > factors. This specificity dictates that one must be cautious in extrapolating the results obtained in rodents to primates. * Nevertheless, despite these various consequences, trans-.genic mice models constitute promising tools with which to study further the various autocrine and paracrine systems Vthat may play a basic role in regulation of follicular growth. Acknowledgments I wish to thank Dr. Marc Antoine Driancourt, Dr. Jan Tesarik, and especially Dr. Dominique De Ziegler for their generous advice and helpful suggestions. H S References 1. Hsueh AJW1986 Paracrine mechanisms involved in granulosa cell ; differentiation. Clin Endocrinol Metab 15: Tonetta SA, DiZerega GS 1989 Intragonadal regulation of follicular maturation. Endocr Rev 10: Giudice LC 1992 Insulin-like growth factors and ovarian follicular ^ development. Endocr Rev 13: Hirshfield AN 1991 Development of follicles in the mammalian y ovary. Int Rev Cytol 124: Gougeon A 1993 Dynamics of human follicular growth: a morphological perspective. In: Adashi EY, Leung PCK (eds) The Ovary. Raven Press, New York, pp Zeleznik AJ 1993 Dynamics of primate follicular growth: a physiologic perspective. In: Adashi EY, Leung PCK (eds) The Ovary. / Raven Press, New York, pp Adashi EY 1994 Endocrinology of the ovary. Hum Reprod 9:815- K Baker TG 1963 A quantitative and cytological study of germ cells in human ovaries. Proc R Soc Lond [Biol] 158: Gougeon A, Chainy GBN 1987 Morphometric studies of small follicles in ovaries of women at different ages. J Reprod Fertil 81: f 10. Koering MJ 1983 Preantral follicle development during the menstrual cycle in the Macaca mulatta ovary. Am J Anat 166: i, 11. Forabosco A, Sforza C, de Pol A, Vizzotto L, Marzona L, Ferrario V1991 Morphometric study of the human neonatal ovary. Anat Rec 231: Block E 1953 A quantitative morphological investigations of the follicular system in newborn female infants. Acta Anat (Basel) 17: ^ 13. Gougeon A, Ecochard R, Thalabard JC 1994 Age-related changes of the population of human ovarian follicles: increase in the dis-», appearance rate of non-growing and early-growing follicles in aging women. Biol Reprod 50: Faddy MJ, Gosden RG, Gougeon A, Richardson SJ, Nelson JF 1992 Accelerated disappearance of ovarian follicles in mid-life: implications for forecasting menopause. Hum Reprod 7: Peters H 1969 The development of the mouse ovary from birth to maturity. Acta Endocrinol (Copenh) 62: Faddy MJ, Jones EC, Edwards RG 1976 An analytical model for ovarian follicle dynamics. J Exp Zool 197: Faddy MJ, Gosden RG 1995 A mathematical model of follicle dynamics in the human ovary. Hum Reprod 10: Meijs Roelofs HMA, Uilenbroek JThJ, Osman P, Welschen R 1973 Serum levels of gonadotrophins and follicular growth in prepubertal rats. In: Peters H (ed) The Development and Maturation of the Ovary and Its Function. International Congress Series. Excerpta Medica, Amsterdam, vol 267: Faddy MJ, Telfer E, Gosden RG 1987 The kinetics of pre-antral follicle development in ovaries of CBA/Ca mice during the first 14 weeks of life. Cell Tissue Kinet 20: Howe E, Lintern-Moore S, Moore GPM, Hawkins J 1978 Ovarian development in hypopituitary Snell dwarf mice. The size and composition of the follicle population. Biol Reprod 19: Halpin DMG, Jones A, Fink G, Charlton HM 1986 Postnatal ovarian follicle development in hypogonadal (hpg) and normal mice and associated changes in the hypothalamic-pituitary ovarian axis. J Reprod Fertil 77: Fagbohun CF, Dada MO, Metcalf JP, Ashiru OA, Blake CA 1990 Blockade of the selective increase in serum follicle-stimulating hormone concentration in immature female rats and its effects on ovarian follicular development. Biol Reprod 42: Peters H, Byskov AG, Lintern-Moore S, Faber M Andersen M 1973 The effect of gonadotrophin on follicle growth initiation in the neonatal mouse ovary. J Reprod Fertil 35: Lintern-Moore S 1977 Initiation of follicular growth in the infant mouse ovary by exogenous gonadotrophin. Biol Reprod 17: Peters H, Braathen B 1973 The effect of unilateral ovariectomy in the neonatal mouse on follicular development. J Endocrinol 56: Arendsen de Wolff-Exalto E 1982 Influence of gonadotrophins on early follicle cell development and early oocyte growth in the immature rat. J Reprod Fertil 66: Motro B, Bernstein A 1993 Dynamic changes in ovarian c-kit and Steel expression during the estrous reproductive cycle. Dev Dynamics 197: Jones EC, Krohn PL 1961 The effect of hypophysectomy on age changes in the ovaries of mice. J Endocrinol 21: Meredith S, Kirkpatrick-Keller D, Butcher R 1986 The effect of food restriction and hypophysectomy on numbers of primordial follicles and concentrations of hormones in rats. Biol Reprod 35: Hirshfield AN 1985 Comparison of granulosa cell proliferation in small follicles of hypophysectomized, prepubertal, and mature rats. Biol Reprod 32: Lintem-Moore S, Moore GPM 1979 The initiation of follicle and oocyte growth in the mouse ovary. Biol Reprod 20: Mariana JC 1978 Follicle cell labelling index as related to primordial follicle oocyte size. Biometric analysis in cycling adult female rat. Ann Biol Anim Biochim Biophys 18: Bukovsky A, Caudle MR, Keenan JA, Wimalasena J, Foster JS, Van Meter SE1995 Quantitative evaluation of the cell cycle-related retinoblastoma protein and localization of Thy-1 differentiation protein and macrophages during follicular development and atresia, and in human corpora lutea. Biol Reprod 52: Gougeon A, Lefevre B, Testart J1992 Morphometric characteristics of the population of the small ovarian follicles in cynomolgus monkey; influence of a gonadotrophin-releasing hormone agonist and gonadotrophins. J Reprod Fertil 95: Krarup T, Pedersen T, Faber M 1969 Regulation of oocyte growth in the mouse ovary. Nature 224: Pedersen T, Peters H 1971 Follicle growth and cell dynamics in the mouse ovary during pregnancy. Fertil Steril 22: Gulyas BJ, Hodgen GD, Tullner WW, Ross GT1977 Effects of fetal or maternal hypophysectomy on endocrine organs and body weight in infant rhesus monkeys (Macaca mulatta) with particular emphasis on oogenesis. Biol Reprod 16:

28 148 GOUGEON Vol. 17, No. 2 i 38. Baker TG, Scrimgeour JB 1980 Development of the gonad in normal and anencephalic human fetuses. J Reprod Fertil 60: Peters H 1979 Some aspects of early follicular development. In: Midgley Jr AR, Sadler Jr WA (eds) Ovarian Follicular Development and Function. Raven Press, New York, pp Neal P, Baker TG 1973 Response of mouse ovaries in vivo and in organ culture to pregnant mare serum gonadotrophins and human chorionic gonadotrophin: examination of critical time interval. J Reprod Fertil 33: O'Shaughnessy PJ, Gray SA 1995 Gonadotropin-dependent and gonadotropin-independent development of inhibin subunit messenger ribonucleic acid levels in the mouse ovary. Endocrinology 136: O'Shaughnessy PJ, Marsh P, Dudley K1994 FSH receptor mrna in the mouse ovary during post-natal development in the normal mouse and in the adult hypogonadal mouse: structure of alternate transcripts. Mol Cell Endocrinol 101: Sokka T, Huhtaniemi Ontogeny of gonadotrophin receptors and gonadotrophin-stimulated cyclic AMP production in the neonatal rat ovary. J Endocrinol 127: Mulheron GW, Quattropani SL, Nolin JM 1989 The ontogeny of immunoreactive endogenous FSH and LH in the rat ovary during early folliculogenesis. Proc Soc Exp Biol Med 190: Ataya K, Tadros M, Ramahi A 1989 Gonadotropin-releasing hormone agonist inhibits physiologic ovarian follicular loss in rats. Acta Endocrinol (Copenh) 121: Lee SJ, Lenton EA, Sexton L, Cooke ID 1988 The effect of age on the cyclical patterns of plasma LH, FSH, oestradiol and progesterone in women with regular menstrual cycles. Hum Reprod 3: Muller P 1963 La dysplasie olfacto genitale chez la femme. Pathol Biol 11: Balboni GC, Vannelli GB, Barni T, Orlando C, Serio M 1987 Transferrin and somatomedin C receptors in the human ovarian follicles. Fertil Steril 48: Zhou J, Chin E, Bondy C 1991 Cellular pattern of insulin-like growth factor-i (IGF-I) and IGF-I receptor gene expression in the developing and mature ovarian follicle. Endocrinology 129: Kasselberg AG, Orth DN, Gray ME, Stahlman MT 1985 Immunocytochemical localization of human epidermal growth factor/ urogastrone in several human tissues. J Histochem Cytochem 33: Chegini N, Williams RS 1992 Immunocytochemical localization of transforming growth factors (TGFs) TGF-a and TGF-/3 in human ovarian tissues. J Clin Endocrinol Metab 74: Maruo T, Ladines-Llave CA, Samoto T, Matsuo H, Manalo AS, Ito H, Mochizuki M 1993 Expression of epidermal growth factor and its receptor in the human ovary during follicular growth and regression. Endocrinology 132: Shull MM, Doetschman T 1994 Transforming growth factor-/31 in reproduction and development. Mol Reprod Dev 39: Maruo T, Hiramatsu S, Otani T, Hayashi M, Mochizuki M 1992 Increase in the expression of thyroid hormone receptors in porcine granulosa cells early in follicular maturation. Acta Endocrinol (Copenh) 127: Li SW, Maruo T, Ladines-Llave CA, Kondo H, Mochizuki M1994 Stage-limited expression of myc oncoprotein in the human ovary during follicular growth, regression and atresia. Endocr J 41: Roy SK 1994 Regulation of ovarian follicular development a review of microscopic studies. Microsc Res Techq 27: Bassett DL 1943 The changes in the vascular pattern of the ovary of the albino rat during the estrous cycle. Am J Anat 73: Jablonska-Shariff A, Fricke PM, Grazul-Bilska AT, Reynolds LP, Redmer DA 1994 Size, number, cellular proliferation, and atresia of gonadotropin-induced follicles in ewes. Biol Reprod 51: Gougeon A 1979 Qualitative changes in medium and large antral follicles in the human ovary during the menstrual cycle. Ann Biol Anim Biochim Biophys 19: Gougeon A, Lefevre B 1983 Evolution of the diameters of the largest healthy and atretic follicles during the human menstrual cycle. J Reprod Fertil 69: McNatty KP, Hillier SG, van den Boogaard AMJ, Trimbos- Kemper TCM, Reichert LE, van Hall EV 1983 Follicular development during the luteal phase of the human menstrual cycle. J Clin Endocrinol Metab 56: Pache TD, Wladimiroff JW, de Jong FH, Hop WC, Fauser BCJM 1990 Growth patterns of non dominant ovarian follicles during the normal menstrual cycle. Fertil Steril 54: Koering MJ, Danforth DR, Hodgen GD 1994 Early follicle growth in the juvenile macaca monkey ovary: the effects of estrogen priming and follicle-stimulating hormone. Biol Reprod 50: Amsterdam A, Rotmensch S 1987 Structure-function relationships during granulosa cell differentiation. Endocr Rev 8: Koering MJ, Goodman AL, Williams RF, Hodgen GD 1982 Gran- ' ulosa cell pyknosis in the dominant follicle of monkeys. Fertil Steril. 37: '" 66. Bomsel Helmreich O, Gougeon A, Thebault A, Saltarelli D,^ Milgrom E, Frydman R, Papiernik E 1979 Healthy and atretic human follicles in the preovulatory phase: differences in evolution of follicular morphology and steroid content of the follicular fluid. J Clin Endocrinol Metab 48: > 67. Balakier H, Stronell RD 1994 Color doppler assessment of folliculogenesis in in vitro fertilization patients. Fertil Steril 62:1211-^ 1216 * 68. DiZerega GS, Hodgen GD 1980 Fluorescence localization of lu- ^ teinizing hormone/human chorionic gonadotropin uptake in the primate ovary. II. Changing distribution during selection of the dominant follicle. J Clin Endocrinol Metab 51: Zeleznik AJ, Schuler HM, Reichert Jr LE 1981 Gonadotropin- t binding sites in the rhesus monkey ovary: role of the vasculature in the selective distribution of human chorionic gonadotropin to * the preovulatory follicle. Endocrinology 109: McClure N, Macpherson AM, Healy DL, Wreford N, Rogers PAW \ 1994 An immunohistochemical study of the vascularization of the human Graafian follicle. Hum Reprod 9: Gougeon A 1982 Rate of follicular growth in the human ovary. In: Rolland R, van Hall EV, Hillier SG, McNatty KP, Schoemaker J (eds) Follicular Maturation and Ovulation. Excerpta Medica, Amsterdam, pp Wiesen JF, Midgley Jr AR 1994 Expression of connexin 43 gap junction messenger ribonucleic acid and protein during follicular K atresia. Biol Reprod 50: Zeleznik AJ, Wildt L, Schuler HM 1980 Characterization of ovar- ^ ian folliculogenesis during the luteal phase of the menstrual cycle in rhesus monkeys using [ 3 H] thymidine autoradiography. Endocrinology 107: Cahill LP, Mauleon P1980 Influence of season, cycle and breed on follicular growth rates in sheep. J Reprod Fertil 58: Goodman AL, Hodgen GS 1983 The ovarian triad of the primate u menstrual cycle. Recent Prog Horm Res 39: Nilsson L, Wikland M, Hamberger L 1982 Recruitment of an ^ ovulatory follicle in the human following follicle-ectomy and luteectomy. Fertil Steril 37: Baird DT, Backstrom T, McNeilly AS, Smith SK, Wathen CG1984 Effect of enucleation of the corpus luteum at different stages of the luteal phase of the human menstrual cycle on subsequent follicular development. J Reprod Fertil 70: ^ 78. Hardy B, Danon D, Eshkol A, Lunenfeld B 1974 Ultrastructural changes in the ovaries of infant mice deprived of endogenous ^ gonadotrophins and after substitution with FSH. J Reprod Fertil 36: Halpin DMG, Charlton HM 1988 Effects of short-term injection of gonadotrophins on ovarian follicle development in hypogonadal (hpg) mice. J Reprod Fertil 82: Santen RJ, Paulsen CA 1973 Hypogonadotropic eunuchoidism. II. ^ Gonadal responsiveness to exogenous gonadotropins. J Clin Endocrinol Metab 36:55-63 / 81. van Weissenbruch MM, Schoemaker HC, Drexhage HA, Schoemaker J 1993 Pharmaco-dynamics of human menopausal > gonadotrophin (HMG) and follicle-stimulating hormone. The importance of the FSH concentration in initiating follicular growth in polycystic ovary-like disease. Hum Reprod 8: Sungurtekin U, Fraser IS, Shearman RP1995 Pregnancy in women * with Kallman's syndrome. Fertil Steril 63: Hodgen GD 1986 Ovarian function for multiple follicle maturation.. Clin Obstet Gynecol 29:

29 April, 1996 OVARIAN FOLLICULAR DEVELOPMENT IN PRIMATES Karnitis VJ, Townson DH, Friedman CI, Danforth DR 1994 Recombinant human follicle-stimulating hormone stimulates multiple follicular growth, but minimal estrogen production in gonadotropin-releasing hormone antagonist-treated monkeys: examining the role of luteinizing hormone in follicular development and steroidogenesis. J Clin Endocrinol Metab 79: Selvaraj N, Mougdal NR 1994 In vivo and in vitro studies on the differential role of luteinizing hormone and follicle-stimulating hormone in regulating follicular function in the bonnet monkey (Macaca radiata) using specific gonadotropin antibodies. Biol Reprod 51: Armstrong DT, Goff AK, Dorrington JH 1979 Regulation of follicular estrogen biosynthesis. In: Midgley Jr AR, Sadler Jr WA (eds) Ovarian Follicular Development and Function. Raven Press, New York, pp Erickson GF 1993 Normal regulation of ovarian androgen production. Semin Reprod Endocrinol 11: Hillier SG, Whitelaw PF, Smyth CD 1994 Follicular oestrogen synthesis: the "two-cell, two-gonadotrophin" model revisited. Mol Cell Endocrinol 100: Nahum R, Thong KJ, Hillier SG 1995 Metabolic regulation of androgen production by human thecal cells in vitro. Hum Reprod 10: Bergh C, Olsson JH, Selleskog U, Hillensjo T 1993 Steroid production in cultured thecal cell obtained from human ovarian follicles. Hum Reprod 8: Hillier SG 1985 Sex steroid metabolism and follicular development in the ovary. Oxf Rev Reprod Biol 7: Sawetawan C, Milewich L, Word PA, Carr BR, Rainey WE 1994 Compartmentalization of type 1 17/3-hydroxysteroid oxidoreductase in the human ovary. Mol Cell Endocrinol 99: Inkster SE, Brodie AMH1991 Expression of aromatase cytochrome P-450 in premenopausal and postmenopausal human ovaries: an immunocytochemical study. J Clin Endocrinol Metab 73: Hillier SG, Reichert Jr LE, van Hall EV 1981 Control of preovulatory estrogen biosynthesis in the human ovary. J Clin Endocrinol Metab 52: Smyth CD, Miro F, Howies CM, Hillier SG 1995 Effect of luteinizing hormone on follicle-stimulating hormone-activated paracrine signalling in rat ovary. Hum Reprod 10: Dennefors BL, Hamberger L, Nilsson L 1983 Influence of human chorionic gonadotropin in vivo on steroid formation and gonadotropin responsiveness of isolated human preovulatory follicular cells. Fertil Steril 39: Peters H 1979 The human ovary in childhood and early maturity. Eur J Obstet Gynecol Reprod Biol 9: Faiman C, Winter JSD 1971 Sex differences in gonadotropin concentration in infancy. Nature 232: Khattab TY, Jequier AM 1979 Serum follicle stimulating hormone levels in human pregnancy. Br J Obstet Gynaecol 86: Govan ADT 1968 The human ovary in early pregnancy. J Endocrinol 40: Ross GT, Lipsett MB 1978 Hormonal correlates of normal and abnormal follicle growth after puberty in humans and other primates. Clin Endocrinol Metab 7: Goldenberg RL, Powell RD, Rosen SW, Marshall JR, Ross GT 1976 Ovarian morphology in women with anosmia and hypogonadotropic hypogonadism. Am J Obstet Gynecol 126: Govan ADT, Black WP 1975 Ovarian morphology in oligomenorrhea. Eur J Obstet Gynecol Reprod Biol 5: Starup J, Sele V, Henriksen B 1971 Amenorrhea associated with increased production of gonadotrophins and a morphologically normal ovarian follicular apparatus. Acta Endocrinol (Copenh) 66: Gougeon A 1984 Influence of cyclic variations in gonadotrophin and steroid hormones on follicular growth in the human ovary. In: de Brux J, Gautray JP (eds) Clinical Pathology of the Endocrine Ovary. MTP Press, Lancaster, UK, pp Testart J, Belaisch-Allart J, Forman R, Gazengel A, Strubb N, Hazout A, Frydman R1989 Influence of different stimulation treatments on oocyte characteristics and in-vitro fertilizing ability. Hum Reprod 4: Chikazawa K, Araki S, Tamada T 1986 Morphological and endocrinological studies on follicular development during the human menstrual cycle. J Clin Endocrinol Metab 62: Roy SK, Treacy BJ 1993 Isolation and long-term culture of human preantral follicles. Fertil Steril 59: Harlow CR, Hillier SG, Hodges JK 1986 Androgen modulation of follicle-stimulating hormone-induced granulosa cell steroidogenesis in the primate ovary. Endocrinology 119: McNatty KP, Makris A, Reinhold VN, De Grazia C, Osathanondh R, Ryan KJ 1979 Metabolism of androstenedione by human ovarian tissues in vitro with particular reference to reductase and aromatase activity. Steroids 34: Sasano H, Okamoto M, Mason JI, Simpson ER, Mendelson CR, Sasano N, Siverberg SG 1989 Immunolocalization of aromatase, 17a-hydroxylase and side-chain cleavage cytochrome P-450 in the human ovary. J Reprod Fertil 85: Tamura T, Kitawaki J, Yamamoto T, Osawa Y, Kominami S, Takemori S, Okada H 1992 Immunohistochemical localization of 17a-hydroxylase/C17-20 lyase and aromatase cytochrome P450 in the human ovary during the menstrual cycle. J Endocrinol 135: Sasano H, Mori T, Sasano N, Nagura H, Mason JI 1990 Immunolocalization of 3/3-hydroxysteroid dehydrogenase in human ovary. J Reprod Fertil 89: Gougeon A, Testart J 1990 Influence of human menopausal gonadotropin on the recruitment of human ovarian follicles. Fertil Steril 54: Brailly S, Gougeon A, Milgrom E, Bomsel-Helmreich O, Papiernik E 1981 Androgens and progestins in the human ovarian follicle: differences in the evolution of preovulatory, healthy nonovulatory, and atretic follicles. J Clin Endocrinol Metab 53: Westergaard L 1985 Follicular atresia in relation to oocyte morphology in non-pregnant and pregnant women. J Reprod Fertil 74: Westergaard L, McNatty KP, Christensen IJ 1985 Steroid concentrations in fluid from human ovarian antral follicles during pregnancy. J Endocrinol 107: , 118. Erickson GF, Hsueh AJW, Quigley ME, Rebar RW, Yen SSC 1979 Functional studies of aromatase activity in human granulosa cells from normal and polycystic ovaries. J Clin Endocrinol Metab 49: Doody KJ, Lorence MC, Mason JI, Simpson ER 1990 Expression of messenger ribonucleic acid species encoding steroidogenic enzymes in human follicles and corpora lutea throughout the menstrual cycle. J Clin Endocrinol Metab McNatty KP 1982 Ovarian follicular development from the onset of luteal regression in humans and sheep. In: Rolland R, van Hall EV, Hillier SG, McNatty KP, Schoemaker J (eds) Follicular Maturation and Ovulation. Excerpta Medica, Amsterdam, pp Gougeon A 1977 Steroid 3/3-ol-dehydrogenase activity in the largest healthy and atretic follicles in the human ovary during the menstrual cycle. Ann Biol Anim Biochim Biophys 17: Richards JS 1993 Gonadotropin-regulated gene expression in the ovary. In: Adashi EY, Leung PCK (eds) The Ovary. Raven Press, New York, pp Kobayashi M, Nakano R, Ooshima A 1990 Immunocytochemical localization of pituitary gonadotrophins and gonadal steroids confirms the "two-cell, two-gonadotrophin" hypothesis of steroidogenesis in the human ovary. J Endocrinol 126: Yamoto M, Shima K, Nakano R 1992 Gonadotropin receptors in human ovarian follicles and corpora lutea throughout the menstrual cycle. Horm Res 37 [Suppl] 1: Steele GL, Leung PCK 1993 Signal transduction mechanisms in ovarian cells. In: Adashi EY, Leung PCK (eds) The Ovary. Raven Press, New York, pp Richards JS 1980 Maturation of ovarian follicles: actions and interactions of pituitary and ovarian hormones on follicular cell differentiation. Physiol Rev 60: Shaw HJ, Hillier SG, Hodges JK 1989 Developmental changes in luteinizing hormone/human chorionic gonadotropin steroidogenic responsiveness in marmoset granulosa cells: effects of follicle-stimulating hormone and androgens. Endocrinology 124: Hillier SG, van Hall EV, van den Bogaard AJM, De Zwart FA,

30 150 GOUGEON Vol. 17, No. 2 Keyzer R 1982 Activation and modulation of the granulosa cell aromatase system: experimental studies with rat and human ovaries. In: Rolland R, van Hall EV, Hillier SG, McNatty KP, Schoemaker J (eds) Follicular Maturation and Ovulation. Excerpta Medica, Amsterdam, pp Sites CK, Patterson K, Jamison CS, Degen SJF, LaBarbera AR1994 follicle-stimulating hormone increases FSH receptor messenger ribonucleic acid while decreasing FSH binding in cultured porcine granulosa cells. Endocrinology 134: Jonassen JA, Bose K, Richards JS 1982 Enhancement and desensitization of hormone-responsive adenylate cyclase in granulosa cells of preantral and antral ovarian follicles: effects of estradiol and follicle-stimulating hormone. Endocrinology 111: Uilenbroek JTJ, Richards JS 1979 Ovarian follicular development during the rat estrous cycle: gonadotropin receptors and follicular responsiveness. Biol Reprod 20: Orly J, Sato G, Erickson GF1980 Serum suppresses the expression of hormonally induced functions in cultured granulosa cells. Cell 20: Savion N, Lui GM, Laherty R, Gospodarowicz D 1981 Factors controlling proliferation and progesterone production by bovine granulosa cells in serum free medium. Endocrinology 109: Gospodarowicz D, 111 CR, Birdwell CR 1977 Effects of fibroblast and epidermal growth factors on ovarian cell proliferation in vitro. I. Characterization of the response of granulosa cells to FGF and EGF. Endocrinology 100: Erickson GF 1983 Primary cultures of ovarian cells in serum-free medium as models of hormone-dependent differentiation. Mol Cell Endocrinol 29: McAllister JM, Mason JI, Byrd W, Trant JM, Waterman MR, Simpson ER 1990 Proliferating human granulosa-lutein cells in long term monolayer culture: expression of aromatase, cholesterol side-chain cleavage, and 3/3-hydroxysteroid dehydrogenase. J Clin Endocrinol Metab 71: Roy SK, Greenwald GS 1988 In vitro effects of follicle-stimulating hormone, luteinizing hormone, and prolactin on follicular deoxyribonucleic acid synthesis in the hamster. Endocrinology 122: Wang X, Roy SK, Greenwald GS 1991 In vitro DNA synthesis by isolated preantral to preovulatory follicles from the cyclic mouse. Biol Reprod 44: Olsson JH, Carlsson B, Hillensjo T 1990 Effect of insulin-like growth factor I on deoxyribonucleic acid synthesis in cultured human granulosa cells. Fertil Steril 54: Angervo M, Koistinen R, Suikkari AM, Seppala M 1991 Insulinlike growth factor binding protein-1 inhibits the DNA amplification induced by insulin-like growth factor I in human granulosa-luteal cells. Hum Reprod 6: Yong EL, Baird DT, Yates R, Reichert Jr LE, Hillier SG 1992 Hormonal regulation of the growth and steroidogenic function of human granulosa cells. J Clin Endocrinol Metab 74: Di Blasio AM, Vigano P, Ferrari A 1994 Insulin-like growth factor-ii stimulates human granulosa-luteal cell proliferation in vitro. Fertil Steril 61: Erickson GF, Garzo VG, Magoffin DA 1989 Insulin-like growth factor-i regulates aromatase activity in human granulosa and human granulosa luteal cells. J Clin Endocrinol Metab 69: Bergh C, Olsson JH, Hillensjo T1991 Effect of insulin-like growth factor I on steroidogenesis in cultured human granulosa cells. Acta Endocrinol (Copenh) 125: Christman GM, Randolph JF, Peegel H Menon KMJ 1991 Differential responsiveness of luteinized human granulosa cells to gonadotropins and insulin-like growth factor I for induction of aromatase activity. Fertil Steril 55: Mason HD, Margara R, Winston RML, Seppala M, Koistinen R, Franks S 1993 Insulin-like growth factor-i (IGF-I) inhibits production of IGF-binding protein-1 while stimulating estradiol secretion in granulosa cells from normal and polycystic human ovaries. J Clin Endocrinol Metab 76: Erickson GF, Garzo VG, Magoffin DA 1991 Progesterone production by human granulosa cells cultured in serum free medium: effects of gonadotrophins and insulin-like growth factor I (IGF-I). Hum Reprod 6: Kubota T, Kamada S, Ohara M, Taguchi M, Sakamoto S, Shimizu Y, Aso T 1993 Insulin-like growth factor II in follicular fluid of the patients with in vitro fertilization and embryo transfer. Fertil Steril 59: Hillier SG, Yong EL, Illingworth PJ, Baird DT, Schwall RH, Mason AJ1991 Effect of recombinant activin on androgen synthesis in cultured human thecal cells. J Clin Endocrinol Metab 72: McAllister JM, Byrd W, Simpson ER 1994 The effects of growth factors and phorbol esters on steroid biosynthesis in isolated human theca interna and granulosa-lutein cells in long term culture. J Clin Endocrinol Metab 79: Bergh C, Carlsson B, Olsson JH, Selleskog U, Hillensjo T 1993 Regulation of androgen production in cultured human thecal cells by insulin-like growth factor 1 and insulin. Fertil Steril 59: Geisthovel F, Moretti-Rojas I, Rojas FJ, Asch RH 1989 Immunoreactive insulin-like growth factor I in human follicular fluid. Hum Reprod 4: Giudice LC, Farrell EM, Pham H, Rosenfeld RG 1990 Identification of insulin-like growth factor-binding protein-3 (IGFBP-3) and IGFBP-2 in human follicular fluid. J Clin Endocrinol Metab 71: Rabinovici J, Dandekar P, Angle MJ, Rosenthal S, Martin MC 1990 Insulin-like growth factor I (IGF-I) levels in follicular fluid from human preovulatory follicles: correlation with serum IGF-I levels. Fertil Steril 54: Hamori M, Blum WF, Torok A, Stehle R, Waibel E, Cledon P, Ranke MB 1991 Insulin-like growth factors and their binding proteins in human follicular fluid. Hum Reprod 6: Jimena P, Castilla JA, Peran F, Molina R, Ramirez JP, Acebal M, Vergara F, Herruzo A 1992 Insulin and insulin-like growth factor I in follicular fluid after induction of ovulation in women undergoing in vitro fertilization. J Reprod Fertil 96: Ramasharma K, Cabrera CM, Li CH 1986 Identification of insulinlike growth factor-ii in human seminal and follicular fluids. Biochem Biophys Res Commun 140: Voutilainen R, Miller WL 1987 Coordinate tropic hormone regulation of mrnas for insulin-like growth factor II and the cholesterol side-chain-cleavage enzyme P450 sec, in human steroidogenic tissues. Proc Natl Acad Sci USA 84: Geisthovel F, Moretti-Rojas I, Asch RH, Rojas FJ 1989 Expression of insulin-like growth factor-ii (IGF-II) messenger ribonucleic acid (mrna), but not IGF-I mrna, in human preovulatory granulosa cells. Hum Reprod 4: Hernandez ER, Hurwitz A, Vera A, Pellicer A, Adashi EY, LeRoith D, Roberts Jr CT 1992 Expression of the genes encoding the insulin-like growth factors and their receptors in the human ovary. J Clin Endocrinol Metab 74: El-Roeiy A, Chen X, Roberts VJ, LeRoith D, Roberts Jr CT, Yen SSC 1993 Expression of insulin-like growth factor-i (IGF-I) and IGF-II and the IGF-I, IGF-II, and insulin receptor genes and localization of the gene products in the human ovary. J Clin Endocrinol Metab 77: Zhou J, Bondy C 1993 Anatomy of the human ovarian insulin-like growth factor system. Biol Reprod 48: Samoto T, Maruo T, Matsuo H, Katayama K, Barnea ER, Mochizuki M 1993 Altered expression of insulin and insulin-like growth factor-i receptors in follicular and stromal compartments of polycystic ovaries. Endocr J 40: Ramasharma K, Li CH 1987 Human pituitary and placental hormones control human insulin-like growth factor II secretion in human granulosa cells. Proc Natl Acad Sci USA 84: Oliver JE, Aitman TJ, Powell JF, Wilson CA, Clayton RN 1989 Insulin-like growth factor-i gene expression in the rat ovary is confined to the granulosa cells of developing follicles. Endocrinology 124: Seppala M, Wahlstrom T, Koskimies AI, Tenhunen A, Rutanen EM, Koistinen R, Huhtaniemi I, Bohn H, Stenman UH 1984 Human preovulatory follicular fluid, luteinizing cells of hyperstimulated preovulatory follicles, and corpus luteum contain placental protein P12. J Clin Endocrinol Metab 58: Hartshorne GM, Bell SC, Waites GT 1990 Binding proteins for insulin-like growth factors in the human ovary: identification, fol-

31 April, 1996 OVARIAN FOLLICULAR DEVELOPMENT IN PRIMATES 151 licular fluid levels and immunohistological localization of the kd type 1 binding protein, IGF-bpI. Hum Reprod 5: Giudice LC, Milki AA, Milkowski DA, El Danasouri I 1991 Human granulosa contain messenger ribonucleic acids encoding insulin-like growth factor-binding proteins (IGFBPs) and secrete IGFBPs in culture. Fertil Steril 56: Holly JMP, Eden JA, Alaghband-zadeh J, Carter GD, Jemmott RC, Cianfarani S, Chard T, Wass JAH 1990 Insulin-like growth factor binding proteins in follicular fluid from normal dominant and cohort follicles, polycystic and multicystic ovaries. Clin Endocrinol (Oxf) 33: Cataldo NA, Giudice LC 1992 Insulin-like growth factor binding protein profiles in human ovarian follicular fluid correlate with functional status. J Clin Endocrinol Metab 74: San Roman GA, Magoffin DA 1993 Insulin-like growth factor binding proteins in healthy and atretic follicles during natural menstrual cycles. J Clin Endocrinol Metab 76: Schuller AGP, Lindenberg-Kortleve DJ, Pache TD, Zwarthoff EC, Fauser BCJM, Drop SLS 1993 Insulin-like growth factor binding protein-2, 28 kda and 24 kda insulin-like growth factor binding protein levels are decreased in fluid of dominant follicles, obtained from normal and polycystic ovaries. Regul Pept 48: Baxter RC, Saunders H 1992 Radioimmunoassay of insulin-like growth factor-binding protein-6 in human serum and other body fluids. J Endocrinol 134: Koistinen R 1990 Human granulosa cells contain insulin-like growth factor-binding protein (IGF BP-1) mrna. Clin Endocrinol (Oxf) 32: Cataldo NA, Woodruff TK, Giudice LC 1993 Regulation of insulinlike growth factor binding protein production by human luteinizing granulosa cells cultured in defined medium. J Clin Endocrinol Metab 76: Jalkanen J, Suikkari AM, Koistinen R, Biitzow R, Ritvos O, Seppala M, Ranta T1989 Regulation of insulin-like growth factorbinding protein-1 production in human granulosa-luteal cells. J Clin Endocrinol Metab 69: Suikkari AM, Jalkanen J, Koistinen R, Biitzow R, Ritvos O, Ranta T, Seppala M 1989 Human granulosa cells synthesize low molecular weight insulin-like growth factor-binding protein. Endocrinology 124: Hamori M, Blum WF, Torok A, Stehle R, Waibel E, Cledon P, Ranke MB 1991 Immunoreactive insulin-like growth factor binding protein-3 in the culture of human luteinized granulosa cells. Acta Endocrinol (Copenh) 124: Elgin RG, Busby Jr WH, Clemmons DR 1987 An insulin-like growth factor (IGF) binding protein enhances the biologic response to IGF-I. Proc Natl Acad Sci USA 84: Mason HD, Willis D, Holly JMP, Cwyf an-hughes SC, Seppala M, Franks S 1992 Inhibitory effects of insulin-like growth factor-binding proteins on steroidogenesis by human granulosa cells in culture. Mol Cell Endocrinol 89:R1-R Bicsak TA, Shimonaka M, Malkowski M, Ling N1990 Insulin-like growth factor-binding protein (IGF-BP) inhibition of granulosa cell function: effects on cyclic adenosine 3', 5'-monophosphate, deoxyribonucleic acid synthesis, and comparison with the effect of an IGF-I antibody. Endocrinology 126: Katz E, Ricciarelli E, Adashi EY 1993 The potential relevance of growth hormone to female reproductive physiology and pathophysiology. Fertil Steril 59: Owen EJ, Shoham Z, Mason BA, Ostergaard H, Jacobs HS 1991 Cotreatment with growth hormone, after pituitary suppression, for ovarian stimulation in in vitro fertilization: a randomized, doubleblind, placebo-control trial. Fertil Steril 56: Ibrahim ZHZ, Matson PL, Buck P, Lieberman BA 1991 The use of biosynthetic human growth hormone to augment ovulation induction with buserelin acetate/human menopausal gonadotropin in women with a poor ovarian response. Fertil Steril 55: Duffy DM, Lindheim SR, Vijod MA, Chang L, Nakamura RM, Lobo RA 1995 Low-dose growth hormone-releasing factor may enhance folliculogenesis in regularly menstruating women: a preliminary study. Fertil Steril 63: Homburg R, West C, Torresani T, Jacobs HS 1990 Cotreatment with human growth hormone and gonadotropins for induction of ovulation: a controlled clinical trial. Fertil Steril 53: Blumenfeld Z, Lunenfeld B 1989 The potentiating effect of growth hormone on follicle stimulation with human menopausal gonadotropin in a panhypopituitary patient. Fertil Steril 52: Shaker AG, Fleming R, Jamieson ME, Yates RWS, Coutts JRT 1992 Absence of effect of adjuvant growth hormone therapy on follicular responses to exogenous gonadotropins in women: normal and poor responders. Fertil Steril 58: Hughes SM, Huang ZH, Morris ID, Matson PL, Buck P, Lieberman BA 1994 A double-blind cross-over controlled study to evaluate the effect of human biosynthetic growth hormone on ovarian stimulation in previous poor responders to in-vitro fertilization. Hum Reprod 9: Dor J, Seidman DS, Amudai E, Bider D, Levran D, Mashiach S 1995 Adjuvant growth hormone therapy in poor responders to in-vitro fertilization: a prospective randomized placebo-controlled double-blind study. Hum Reprod 10: Hughes JN, Torresani T, Herve F, Martin-Point B, Tamboise A, Santarelli J 1991 Interest of growth hormone-releasing hormone administration for improvement of ovarian responsiveness to gonadotropins in poor responder women. Fertil Steril 55: Shaker AG, Fleming R, Jamieson ME, Yates RWS, Coutts JRT 1992 Ovarian stimulation in an infertile patient with growth hormone-deficient Oliver-Mcfarlane syndrome. Hum Reprod 9: Carlsson B, Bergh C, Bentham J, Olsson JH, Norman MR, Billig H, Roos P, Hillensjo T 1992 Expression of functional growth hormone receptors in human granulosa cells. Hum Reprod 7: Sharar FI, Nieman LK 1994 Identification and cellular localization of growth hormone receptor gene expression in the human ovary. J Clin Endocrinol Metab 79: Tamura M, Sasano H, Suzuki T, Fukaya T, Watanabe T, Aoki H, Nagura H, Yajima A 1994 Immunohistochemical localization of growth hormone receptor in cyclic human ovaries. Hum Reprod 9: Menashe Y, Sack J, Mashiach S 1991 Spontaneous pregnancies in two women with Laron-type dwarfism: are growth hormone and circulating insulin-like growth factor mandatory for induction of ovulation? Hum Reprod 5: Dor J, Ben-Shlomo I, Lunenfeld B, Pariente C, Levran D, Karasik A, Seppala M, Mashiach S 1992 Insulin-like growth factor-i (IGF-I) may not be essential for ovarian follicular development: evidence for IGF-I deficiency. J Clin Endocrinol Metab 74: Mulheron GW, Schomberg DW 1993 The intraovarian transforming growth factor system. In: Adashi EY, Leung PCK (eds) The Ovary. Raven Press, New York, pp Gospodarowicz D, Bialecki H 1979 Fibroblast and epidermal growth factors are mitogenic agents for cultured granulosa cells of rodent, porcine, and human origin. Endocrinology 104: TapanainenJ, Leinonen PJ, Tapanainen P, Yamamoto M, Jaffe RB 1987 Regulation of human granulosa-luteal cell progesterone production and proliferation by gonadotropins and growth factors. Fertil Steril 48: Tilly JL, LaPolt PS, Hsueh AJW 1992 Hormonal regulation of follicle-stimulating hormone receptor messenger ribonucleic acid levels in cultured rat granulosa cells. Endocrinology 130: Hasegawa Y, Miyamoto K, Abe Y, Nakamura T, Sugino H, Eto Y, Shibai H, Igarashi M 1988 Induction of follicle-stimulating hormone receptor by erythroid differentiation factor on rat granulosa cell. Biochem Biophys Res Commun 156: Mondschein JS, Schomberg DW 1981 Growth factors modulate gonadotropin receptor induction in granulosa cell cultures. Science 211: Mondschein JS, May JV, Gunn EB, Schomberg DW 1981 Modulation by epidermal growth factor of LH/hCG receptor binding in porcine granulosa cell cultures: implications for follicle development and atresia. In: Schwartz NB, Hunzicker-Dunn M (eds) Dynamics of Ovarian Function. Raven Press, New York, pp Mason HD, Margara R, Winston RML, Beard RW, Reed MJ, Franks S 1990 Inhibition of oestradiol production by epidermal growth factor in human granulosa cells of normal and polycystic ovaries. Clin Endocrinol (Oxf) 33:

32 152 GOUGEON Vol. 17, No Steinkampf MP, Mendelson CR, Simpson ER 1988 Effects of epidermal growth factor and insulin-like growth factor I on the levels of mrna encoding aromatase cytochrome P-450 of human ovarian granulosa cells. Mol Cell Endocrinol 59: Franchimont P, Hazee-Hagelstein MT, Charlet-Renard CH, Jasper JM1986 Effect of mouse epidermal growth factor on DNA and protein synthesis, progesterone and inhibin production by bovine granulosa cell cultures. Acta Endocrinol (Copenh) 111: Erickson GF, Case E 1983 Epidermal growth factor antagonizes ovarian theca-interstitial cytodifferentiation. Mol Cell Endocrinol 31: Richardson MC, Gadd SC, Masson GM 1989 Augmentation by epidermal growth factor of basal and stimulated progesterone production by human luteinized granulosa cells. J Endocrinol 121: Angervo M, Koistinen R, Seppala M 1992 Epidermal growth factor stimulates production of insulin-like growth factor-binding protein-1 in human granulosa-luteal cells. J Endocrinol 134: Westergaard LG Andersen CY 1989 Epidermal growth factor (EGF) in human preovulatory follicles. Hum Reprod 4: Westergaard LG, Andersen CY, Byskov AG 1990 Epidermal growth factor in small antral ovarian follicles of pregnant women. J Endocrinol 127: Franks S, Mason HD, Poison DW, Winston RML, Margara R, Reed MJ 1988 Mechanism and management of ovulatory failure in women with polycystic ovary syndrome. Hum Reprod 3: Mason HD, Carr L, Leake R, Franks S 1995 Production of transforming growth factor-a by normal and polycystic ovaries. J Clin Endocrinol Metab 80: Baird A, Esch F, Mormede P, Ueno N, Ling N, Bohlen P, Ying SY, Wehrenberg WB, Guillemin R 1986 Molecular characterization of fibroblast growth factor: distribution and biological activities in various tissues. Recent Prog Horm Res 42: Adashi EY, Resnick CE, Croft CS, May JV, Gospodarowicz D1988 Basic fibroblast growth factor as a regulator of ovarian granulosa cell differentiation: a novel non-mitogenic role. Mol Cell Endocrinol 55: Koos RD 1993 Ovarian angiogenesis. In: Adashi EY, Leung PCK (eds) The Ovary. Raven Press, New York, pp Yeh J, Osathanondh R 1993 Expression of messenger ribonucleic acids encoding for basic fibroblast growth factor (FGF) and alternatively spliced FGF receptor in human fetal ovary and uterus. J Clin Endocrinol Metab 77: Watson R, Anthony F, Pickett M, Lambden P, Masson GM, Thomas EJ 1992 Reverse transcription with nested polymerase chain reaction shows expression of basic fibroblast growth factor transcripts in human granulosa and cumulus cells from in vitro fertilization patients. Biochem Biophys Res Commun 187: Di Blasio AM, Vigano OP, Cremonesi L, Carniti C, Ferrari M, Ferrari A 1993 Expression of the genes encoding basic fibroblast growth factor and its receptor in human granulosa cells. Mol Cell Endocrinol 96:R7-R Zhiwen Z, Findlay JK, Carson RD, Herington AC, Burger HG 1988 Transforming growth factor-j3 enhances basal and FSHstimulated inhibin production by rat granulosa cells in vitro. Mol Cell Endocrinol 58: Gitay-Goren H, Kim I-C, Miggans ST, Schomberg DW 1993 Transforming growth factor-beta modulates gonadotropin receptor expression in porcine and rat granulosa cells differently. Biol Reprod 48: Chegini N, Flanders KC 1992 Presence of transforming growth factor-/3 and their selective cellular localization in human ovarian tissue of various reproductive stages. Endocrinology 130: Mulheron GW, Bossert NL, Lapp JA, Walmer DK, Schomberg DW 1992 Human granulosa-luteal and cumulus cells express transforming growth factors-beta type 1 and type 2 mrna. J Clin Endocrinol Metab 74: Gangrade BK, May JV 1990 The production of transforming growth factor-/? in the porcine ovary and its secretion in vitro. Endocrinology 127: Findlay JK, Xiao S, Shukovski L, Michel U 1993 Novel peptides in ovarian physiology: inhibin, activin, and follistatin. In: Adashi EY, Leung PCK (eds) The Ovary. Raven Press, New York, pp Miro F, Hillier SG 1992 Relative effects of activin and inhibin on steroid hormone synthesis in primate granulosa cells. J Clin Endocrinol Metab 75: Hillier SG, Yong EL, Illingworth PJ, Baird DT, Schwall RH, Mason AJ 1991 Effect of recombinant inhibin on androgen synthesis in cultured human thecal cells. Mol Cell Endocrinol 75:R1-R Rabinovici J, Spencer SJ, Jaffe RB 1990 Recombinant human activin-a promotes proliferation of human luteinized preovulatory granulosa cells in vitro. J Clin Endocrinol Metab 71: Li W, Ho-Yuen B, Leung PCK 1992 Inhibition of progestin accumulation by activin-a in human granulosa cells. J Clin Endocrinol Metab 75: Rabinovici J, Spencer SJ, Doldi N, Goldsmith PC, Schwall R, Jaffe RB 1992 Activin-A as an intraovarian modulator: actions, localization, and regulation of the intact dimer in human ovarian cells. J Clin Invest 89: Xiao S, Robertson DM, Findlay JK 1992 Effects of activin and follicle-stimulating hormone (FSH)-suppressing protein/ follistatin on FSH receptors and differentiation of cultured rat granulosa cells. Endocrinology 131: Hsueh AJW, Dahl KD, Vaughan J, Tucker E, Rivier J, Bardin CW, Vale W 1987 Heterodimers and homodimers of inhibin subunits have different paracrine action in the modulation of luteinizing hormone-stimulated androgen biosynthesis. Proc Natl Acad Sci USA 84: Hutchinson LA, Findlay JK, de Vos FL, Robertson DM 1987 Effects of bovine inhibin, transforming growth factor-j3 and bovine activin-a on granulosa cell differentiation. Biochem Biophys Res Commun 146: Li W, Khorasheh S, Yuen BH, Ling N, Leung PCK 1993 Stimulation of progesterone secretion by recombinant follistatin-288 in human granulosa cells. Endocrinology 132: Matzuk MM, Finegold MJ, Su JGJ, Hsueh AJW, Bradley A 1992 a-lnhibin is a tumour-suppressor gene with gonadal specificity in mice. Nature 360: Li R, Phillips DM, Mather JP 1995 Activin promotes ovarian follicle development in vitro. Endocrinology 136: Cameron VA, Nishimura E, Mathews LS, Lewis KA, Sawchenko PE, Vale WW 1994 Hybridization histochemical localization of activin receptor subtypes in rat brain, pituitary, ovary, and testis. Endocrinology 134: Woodruff TK, Krummen L, McCray G, Mather JP 1993 In situ ligand binding of recombinant human [ 125 I] activin -A and recombinant human [ 125 I] inhibin-a to the adult rat ovary. Endocrinology 133: Roberts VJ, Barth S, El-Roeiy A, Yen SSC 1994 Expression of inhibin/ activin system messenger ribonucleic acids and proteins in ovarian follicles from women with polycystic ovarian syndrome. J Clin Endocrinol Metab 79: Nakamura T, Takio K, Eto Y, Shibai H, Titani K, Sugino H 1990 Activin-binding protein from rat ovary is follistatin. Science 247: Schneyer AL, O'Neil DA, Crowley Jr WF 1992 Activin-binding proteins in human serum and follicular fluid. J Clin Endocrinol Metab 74: Hillier SG, Wickings EJ, Illingworth PJ, Yong EL, Reichert Jr LE, Baird DT, McNeilly AS 1991 Control of immunoreactive inhibin production by human granulosa cells. Clin Endocrinol (Oxf) 35: Hillier SG, Wickings EJ, Saunders PTK, Dixson AF, Shimasaki S, Swanston IA, Reichert Jr LE, McNeilly AS 1989 Control of inhibin production by primate granulosa cells. J Endocrinol 123: Schwall RH, Mason AJ, Wilcox JN, Bassett SG, Zeleznik AJ 1990 Localization of inhibin/ activin subunit mrnas within the primate ovary. Mol Endocrinol 4: Fraser HM, Lunn SF, Cowen GM, Saunders PTK 1993 Localization of inhibin/activin subunit mrnas during the luteal phase in the primate ovary. J Mol Endocrinol 10: Marrs RP, Lobo R, Campeau JD, Nakamura RM, Brown J, Ujita EL, DiZerega GS1984 Correlation of human follicular fluid inhibin

33 April, 1996 OVARIAN FOLLICULAR DEVELOPMENT IN PRIMATES 153 activity with spontaneous and induced follicle maturation. J Clin Endocrinol Metab 58: Yamoto M, Minami S, Nakano R, Kobayashi M 1992 Immunohistochemical localization of inhibin/activin subunits in human ovarian follicles during the menstrual cycle. J Clin Endocrinol Metab 74: Yamoto M, Minami S, Nakano R 1993 Immunohistochemical localization of inhibin subunits in polycysric ovary. J Clin Endocrinol Metab 77: Roberts VJ, Barth S, El-Roeiy A, Yen SSC 1993 Expression of inhibin/activin subunits and follistatin messenger ribonucleic acids and proteins in ovarian follicles and the corpus luteum during the human menstrual cycle. J Clin Endocrinol Metab 77: DePaolo LV, Bicsak TA, Erickson GF, Shimasaki S, Ling N 1991 Follistatin and activin: a potential intrinsic regulator system within diverse tissues. Proc Soc Exp Biol Med 198: Adashi EY 1990 The potential relevance of cytokines to ovarian physiology: the emerging role of resident ovarian cells of the white blood cell series. Endocr Rev 11: Ben-Rafael Z, Orvieto R 1992 Cytokines - involvement in reproduction. Fertil Steril 58: Terranova PF, Sancho-Tello M, Hunter VJ 1993 Tumor necrosis factor-a and ovarian function. In: Adashi EY, Leung PCK (eds) The Ovary. Raven Press, New York, pp Wang LJ, Brannstrom M, Robertson SA, Norman RJ 1992 Tumor necrosis factor a in the human ovary: presence in follicular fluid and effects on cell proliferation and prostaglandin production. Fertil Steril 58: Yan Z, Hunter V, Weed J, Hutchison S, Lyles R, Terranova P 1993 Tumor necrosis factor-a alters steroidogenesis and stimulates proliferation of human ovarian granulosal cells in vitro. Fertil Steril. 59: Fukuoka M, Yasuda K, Fujiwara H, Kanzaki H, Mori T 1992 Interactions between interferon y, tumour necrosis factor a, and interleukin-1 in modulating progesterone and oestradiol production by human luteinized granulosa cells in culture. Hum Reprod 7: Zolti M, Meiron R, Shemesh M, Wollach D, Mashiach S, Shore L, Ben Rafael Z 1990 Granulosa cells as a source and target organ for tumor necrosis factor-a. FEBS Lett 261: Wang L, Robertson S, Seamark RF, Norman RJ 1991 Lymphokines, including interleukin-2, alter gonadotropin-stimulated progesterone production and proliferation of human granulosa-luteal cells in vitro. J Clin Endocrinol Metab 72: Barak V, Yanai P, Treves AJ, Roisman I, Simon A, Laufer N 1992 Interleukin-1: local production and modulation of human granulosa luteal cells steroidogenesis. Fertil Steril 58: Fukuoka M, Yasuda K, Emi N, Fujiwara H, Iwai M, Takakura K, Kanzaki H, Mori T1992 Cytokine modulation of progesterone and estradiol secretion in cultures of luteinized human granulosa cells. J Clin Endocrinol Metab 75: Kauppila A, Cantell K, Janne O, Kokko E, Vihko R 1982 Serum sex steroid and peptide hormone concentrations, and endometrial estrogen and progestin receptor levels during administration of human leukocyte interferon. Int J Cancer 29: Hurwitz A, Loukides J, Ricciarelli E, Botero L, Katz E, McAllister JM, Garcia JE, Rohan R, Adashi EY, Hernandez ER 1992 Human intraovarian interleukin-1 (IL-1) system: highly compartmentalized and hormonally dependent regulation of the genes encoding IL-1, its receptor, and its receptor antagonist. J Clin Invest 89: Piquette GN, Simon C, El-Danasouri I, Frances A, Polan ML 1994 Gene regulation of interleukin-1 /3, interleukin-1 receptor type 1, and plasminogen activator inhibitor-1 and -2 in human granulosaluteal cells. Fertil Steril 62: Roby KB, Weed J, Lyles R, Terranova PF 1990 Immunological evidence for a human ovarian tumor necrosis factor-a. J Clin Endocrinol Metab 71: Grasso G, Muscettola M, Traina V, Causio F, Fanizza G, Cagnazzo G 1988 Presence of interferons in human follicular fluid after ovarian hypersrimulation for in vitro fertilization. Med Sci Res 16: Khan SA, Schmid K, Hallin P, Di Pauli R, De Geyter C, Nieschlag E 1988 Human testis cytosol and ovarian follicular fluid contain high amounts of interleukin-1-like factor(s). Mol Cell Endocrinol 58: Wang LJ, Norman RJ 1992 Concentrations of immunoreactive interleukin-1 and interleukin-2 in human preovulatory follicular fluid. Hum Reprod 7: Machelon V, Emilie D, Lefevre A, Nome F, Durand-Gasselin I, Testart J 1994 Interleukin-6 biosynthesis in human preovulatory follicles: some of its potential roles at ovulation. J Clin Endocrinol Metab 79: Buyalos RP, Watson JM, Martinez-Maza O 1992 Detection of interleukin-6 in human follicular fluid. Fertil Steril 57: Brannstrom M, Pascoe V, Norman RJ, McClure N 1994 Localization of leucocyte subsets in the follicle wall and in the corpus luteum throughout the human menstrual cycle. Fertil Steril 61: Loukides JA, Loy RA, Edwards R, Honig J, Visintin I, Polan ML 1990 Human follicular fluids contain tissue macrophages. J Clin Endocrinol Metab 71: Seow WK, Thong YH, Waters MJ, Walters M, Cummins JM 1988 Isolation of a chemotactic protein for neutrophils from human ovarian follicular fluid. Int Arch Allergy Appl Immunol 86: Espey L 1994 Current status of the hypothesis that mammalian ovulation is comparable to an inflammatory reaction. Biol Reprod 50: Tsafriri A, Braw RH 1984 Experimental approaches to atresia in mammals. Oxf Rev Reprod Biol 6: Hughes FM, Gorospe WC 1991 Biochemical identification of apoptosis (programmed cell death) in granulosa cells: evidence for a potential mechanism underlying follicular atresia. Endocrinology 129: Billig H, Furuta I, Hsueh AJW 1993 Estrogens inhibit and androgens enhance ovarian granulosa cell apoptosis. Endocrinology 133: Hurwitz A, Adashi EY 1993 Ovarian follicular atresia as an apoptotic process. In: Adashi EY, Leung PCK (eds) The Ovary. Raven Press, New York, pp Gougeon A, Testart J 1986 Germinal vesicle breakdown in oocytes of human atretic follicles during the menstrual cycle. J Reprod Fertil 78: McNatty KP, Moore Smith D, Makris A, Osathanondh R, Ryan KJ 1979 The microenvironment of the human antral follicle: interrelationships among the steroid levels in antral fluid, the population of granulosa cells, and the status of the oocyte in vivo and in vitro. J Clin Endocrinol Metab 49: Hillier SG, van den Boogaard AMS, Reichert Jr LE, van Hall EV 1980 Intraovarian sex steroid hormone interactions and the regulation of follicular maturation: aromatization of androgens by human granulosa cells in vitro. J Clin Endocrinol Metab 50: Shima K, Kitayama S, Nakano R 1987 Gonadotropin binding sites in human ovarian follicles and corpora lutea during the menstrual cycle. Obstet Gynecol 69: Deane HW, Lobel BL, Romney SL 1962 Enzymic histochemistry of normal ovaries of the menstrual cycle, pregnancy and early puerperium. Am J Obstet Gynecol 83: Breitenecker G, Friedrich F, Kemeter P1978 Further investigations on the maturation and degeneration of human ovarian follicles and their oocytes. Fertil Steril 29: Behrman HR, Riley JCM, Aten RF 1993 Reactive oxygen species and ovarian function. In: Adashi EY, Leung PCK (eds) The Ovary. Raven Press, New York, pp McConkey DJ, Hartzell P, Nicotera P, Wyllie AH, Orrenius S1988 Stimulation of endogenous endonuclease activity in hepatocytes exposed to oxidative stress. Toxicol Lett 42: Zeleznik AJ, Ihrig LL, Basseti SG 1989 Developmental expression of Ca + + /Mg ++ -dependent endonuclease activity in rat granulosa and luteal cells. Endocrinology 125: Tilly JL, Tilly KI 1995 Inhibitors of oxidative stress mimic the ability of follicle-stimulating hormone to suppress apoptosis in cultured rat ovarian follicles. Endocrinology 136: Tilly KI, Banerjee S, Banerjee PP, Tilly JL 1995 Expression of the p53 and Wilms' tumor suppressor genes in the rat ovary: gonadotropin repression in vivo and immunohistochemical localization

34 154 GOUGEON Vol. 17, No. 2 of nuclear p53 protein to apoptotic granulosa cells of atretic follicles. Endocrinology 136: Tilly JL, Billig H, Kowalski KI, Hsueh AJW 1992 Epidermal growth factor and basic fibroblast growth factor suppress the spontaneous onset of apoptosis in cultured rat ovarian granulosa cells and follicles by a tyrosine kinase-dependent mechanism. Mol Endocrinol 6: Tilly JL, Tilly KI, Kenton ML, Johnson AL 1995 Expression of members of the Bcl-2 gene family in the immature rat ovary: equine chorionic gonadotropin-mediated inhibition of granulosa cell apoptosis is associated with decreased Bax and constitutive Bcl-2 and Bcl-xlong messenger ribonucleic acid levels. Endocrinology 136: Farookhi R 1981 Atresia: an hypothesis. In: Schwartz NB, Hunzicker-Dunn M (eds) Dynamics of Ovarian Function. Raven Press, New York, pp Billig H, Furuta I, Hsueh AJW 1994 GnRH directly induces apoptotic cell death in the rat ovary: biochemical and in situ detection of DNA fragmentation in granulosa cells. Endocrinology 134: Erickson GF, Li DL, Sadrkhanloo R, Liu XJ, Shimasaki S, Ling N 1994 Extrapituitary actions of gonadotropin-releasing hormone: stimulation of insulin-like growth factor-binding protein-4 and atresia. Endocrinology 134: Wang J, Baimbridge KG, Leung PCK 1989 Changes in cytosolic free calcium ion concentrations in individual rat granulosa cells: effects of luteinizing hormone-releasing hormone. Endocrinology 124: Kitzman PH, Hutz RJ 1992 In vitro effects of angiotensin-11 on steroid production by hamster ovarian follicles and on ultrastructure of the theca-interna. Cell Tissue Res 268: Mukhopadhyay AK, Holstein K, Szkudlinski M, Brunswig- Spickenheier B, Leidenberger FA 1991 The relationship between prorenin levels in follicular fluid and follicular atresia in bovine ovaries. Endocrinology 129: Latouche J, Crumeyrolle-Arias M, Jordan D, Kopp N, Augendre- Ferrante B, Cedard L, Haour F 1989 GnRH receptors in human granulosa cells: anatomical localization and characterization by autoradiographic study. Endocrinology 125: Peng C, Fan NC, Ligier M, Vaananen J, Leung PCK 1994 Expression and regulation of gonadotropin-releasing hormone (GnRH) and GnRH receptor messenger ribonucleic acids in human granulosa-luteal cells. Endocrinology 135: Minaretzis D, Jakubowski M, Mortola JF, Pavlou SN 1995 Gonadotropin-releasing hormone receptor gene expression in human ovary and granulosa-lutein cells. J Clin Endocrinol Metab 80: Palumbo A, Pourmotabbed G, Carcangiu ML, Andrade-Gordon P, Roa L, DeCherney A, Naftolin F 1993 Immunohistochemical localization or renin and angiotensin in the ovary: comparison between normal women and patients with histologically proven polycystic ovarian disease. Fertil Steril 60: Harlow CR, Shaw HJ, Hillier SG, Hodges JK 1986 Factors influencing follicle-stimulating hormone-responsive steroidogenesis in marmoset granulosa cells: effects of androgens and the stage of follicular maturity. Endocrinology 122: Erickson GF, Nakatani A, Ling N, Shimasaki S 1992 Localization of insulin-like growth factor-binding protein-4 messenger ribonucleic acid in the rat ovary. Endocrinology 130: Cataldo NA, Giudice LC 1992 Follicular fluid insulin-like growth factor binding protein profiles in polycystic ovary syndrome. J Clin Endocrinol Metab 74: Ryan KJ1981 Follicular atresia some speculations of biochemical markers and mechanisms. In: Schwartz NB, Hunzicker-Dunn M (eds) Dynamics of Ovarian Function. Raven Press, New York, pp Pepperell JR, Nemeth G, Palumbo A, Naftolin F 1993 The intraovarian renin-angiotensin system. In: Adashi EY, Leung PCK (eds) The Ovary. Raven Press, New York, pp Stirling D, Magness RR, Stone R, Waterman MR, Simpson ER 1990 Angiotensin II inhibits luteinizing hormone-stimulated cholesterol side chain cleavage expression and stimulates basic fibroblast growth factor expression in bovine luteal cells in primary culture. J Biol Chem 265: McClure N, Macpherson AM, Abberton KM, Healy DL, Rogers PAW 1993 Human follicular fluid maturity and endothelial cell mitogenesis. Hum Reprod 8: Katz E, Goodman AL, Garcia J, Rone JD 1992 Evidence for the production of a nonsteroidal endotheliotropic factor by human granulosa cells in culture. Fertil Steril 57: Ravindranath N, Little-Ihrig L, Phillips HS, Ferrara N, Zeleznik AJ 1992 Vascular endothelial growth factor messenger ribonucleic acid expression in the primate ovary. Endocrinology 131: Schreiber AB, Winkler ME, Derynck R1986 Transforming growth factor-a: a more potent angiogenic mediator than epidermal growth factor. Science 232: Erickson GF, Magoffin DA, Dyer CA, Hofeditz C 1985 The ovarian androgen producing cells: a review of structure/function relationships. Endocr Rev 6: Magoffin DA, Weitsman SR 1994 Insulin-like growth factor-1 regulation of luteinizing hormone (LH) receptor messenger ribonucleic acid expression and LH-stimulated signal transduction in rat ovarian theca-interstitial cells. Biol Reprod 51: Schwall R, Erickson GF 1981 Functional and morphological changes in rat theca cells during atresia. In: Schwartz NB, Hunzicker-Dunn M (eds) Dynamics of Ovarian Function. Raven Press, New York, pp Oakberg EF 1979 Follicular growth and atresia in the mouse. In Vitro 15: Greenwald GS, Terranova PF 1988 Follicular selection and its control. In: Knobil E, Neil JD (eds) The Physiology of Reproduction. Raven Press, New York, vol 1: Roy SK, Greenwald GS 1990 Immunohistochemical localization of epidermal growth factor-like activity in the hamster ovary with a polyclonal antibody. Endocrinology 126: Jones PBC, Welsh Jr TH, Hsueh AJW 1982 Regulation of ovarian progestin production by epidermal growth factor in cultured rat granulosa cells. J Biol Chem 257: LaPolt PS, Tilly JL, Aihara T, Nishimori K, Hsueh AJW 1992 Gonadotropin-induced up- and down-regulation of ovarian follicle-stimulating hormone (FSH) receptor gene expression in immature rats: effects of pregnant mare's serum gonadotropin, human chorionic gonadorropin, and recombinant FSH. Endocrinology 130: Sato A, Bo M, Maruo T, Yoshida S, Mochizuki M 1994 Stagespecific expression of c-myc messenger ribonucleic acid in porcine granulosa cells early in follicular growth. Eur J Endocrinol 131: Knecht M, Catt KJ 1983 Modulation of camp-mediated differentiation in ovarian granulosa cells by epidermal growth factor and platelet-derived growth factor. J Biol Chem 258: Zachow RJ, Tash JS, Terranova PF 1992 Tumor necrosis factor-a induces clustering in ovarian theca-interstitial cells in vitro. Endocrinology 131: Karakji EG, Tsang BK 1995 Tumor necrosis factor alpha inhibits rat granulosa cell plasminogen activator activity in vitro during follicular development. Biol Reprod 52: Sugino H, Nakamura T, Hasegawa Y, Miyamoto K, Igarashi M, Eto Y, Shibai H, Titani K 1988 Identification of a specific receptor for erythroid differentiation factor on follicular granulosa cell. J Biol Chem 263: Kranzfelder D, Maurer-Schultze B 1989 Development of the perifollicular capillary network. Autoradiographic and morphometric studies in the rabbit ovary. Eur J Obstet Gynecol Reprod Biol 30: May JV, Frost JP, Schomberg DW 1988 Differential effects of epidermal growth factor, somatomedin-c/ insulin-like growth factor-1, and transforming growth factor-/? on porcine granulosa cell deoxyribonucleic acid synthesis and cell proliferation. Endocrinology 123: Hirshfield AN 1986 Patterns of 3 H thymidine incorporation differ in immature rats and mature cycling rats. Biol Reprod 34: Liu XJ, Malkowski M, Guo Y, Erickson GF, Shimasaki S, Ling N 1993 Development of specific antibodies to rat insulin-like growth factor-binding proteins (IGFBP-2 to -6): analysis of IGFBP production by rat granulosa cells. Endocrinology 132: Soules MR, Steiner RA, Clifton DK, Cohen NL, Aksel S, Bremner WJ1984 Progesterone modulation of pulsatile luteinizing hormone secretion in normal women. J Clin Endocrinol Metab 58:

35 April, 1996 OVARIAN FOLLICULAR DEVELOPMENT IN PRIMATES Boland NI, Gosden RG 1994 Clonal analysis of chimaeric mouse ovaries using DNA in situ hybridization. J Reprod Fertil 100: Telfer E, Ansell JD, Taylor H, Gosden RG 1988 The number of clonal precursors of the follicular epithelium in the mouse ovary. J Reprod Fertil 84: Sugawara M, DePaolo L, Nakatani A, DiMarzo S, Ling N 1990 Radioimmunoassay of follistatin: application for in vitro fertilization procedures. J Clin Endocrinol Metab 71: Nakatani A, Shimasaki S, DePaolo LV, Erickson GF, Ling N 1991 Cyclic changes in follistatin messenger ribonucleic acid and its protein in the rat ovary during the estrous cycle. Endocrinology 129: Michel U, McMaster JW, Findlay JK 1992 Regulation of steady state follistatin messenger RNA levels in rat granulosa cells in vitro. J Mol Endocrinol 9: Xiao S, Findlay JK 1991 Interactions between activin and FSH suppressing protein and their mechanisms of action on cultured rat granulosa cells. Mol Cell Endocrinol 79: McLachlan RI, Cohen NL, Dahl KD, Bremner WJ, Soules MR 1990 Serum inhibin levels during the periovulatory interval in normal women: relationships with sex steroid and gonadotrophin levels. Clin Endocrinol (Oxf) 32: Caruso A, Fortini A, Fulghesu AM, Pistilli E, Cucinelli F, Lanzone A, Marcuso S 1993 Ovarian sensitivity to follicle-stimulating hormone during the follicular phase of the human menstrual cycle and in patients with polycystic ovarian syndrome. Fertil Steril 59: Hone K, Takakura K, Fujiwara H, Suginami H, Liao S, Mori T1992 Immunohistochemical localization of androgen receptor in the human ovary throughout the menstrual cycle in relation to oestrogen and progesterone receptor expression. Hum Reprod 7: Suzuki T, Sasano H, Kimura N, Tamura M, Fukaya T, Yajima A, Nagura H 1994 Immunohistochemical distribution of progesterone, androgen and oestrogen receptors in the human ovary during the menstrual cycle: relationship to expression of steroidogenic enzymes. Hum Reprod 9: Iwai T, Nanbu Y, Iwai M, Taii S, Fujii S, Mori T 1990 Immunohistochemical localization of oestrogen receptors and progesterone receptors in the human ovary throughout the menstrual cycle. Virchows Arch [A] 417: Billiar RB, Loukides JA, Miller MM 1992 Evidence for the presence of the estrogen receptor in the ovary of the baboon (Papio anubis). J Clin Endocrinol Metab 75: Hurst BS, Zilberstein M, Chou JY, Litman BL, Stephens J, Leslie KK 1995 Estrogen receptors are present in human granulosa cells. J Clin Endocrinol Metab 80: Rabinovici J, Blankstein J, Goldman B, Rudak E, Dor Y, Pariente C, Geier A, Lunenfeld B, Mashiach S 1989 In vitro fertilization and primary embryonic cleavage are possible in 17a-hydroxylase deficiency despite extremely low intrafollicular 17j3-estradiol. J Clin Endocrinol Metab 68: Lavranos TC, Rodgers HF, Bertoncello I, Rodgers RJ 1994 Anchorage-independent culture of bovine granulosa cells: the effects of basic fibroblast growth factor, dibutyryl camp on cell division, differentiation. Exp Cell Res 211: Hodgen GD 1982 The dominant ovarian follicle. Fertil Steril 38: DiZerega GS, Hodgen GD 1980 The primate ovarian cycle suppression of human menopausal gonadotropin-induced follicular growth in the presence of the dominant follicle. J Clin Endocrinol Metab 50: DiZerega GS, Tonetta SA, Westhof G 1987 A postulated role for naturally occurring aromatase inhibitors in follicle selection. J Steroid Biochem 27: Zeleznik AJ, Hutchison JS, Schuler HM 1985 Interference with the gonadotropin-suppressing actions of estradiol in macaques overrides the selection of a single preovulatory follicle. Endocrinology 117: Zeleznik AJ, Kubick CJ 1986 Ovarian responses in macaques to pulsatile infusion of follicle-stimulating hormone (FSH) and luteinizing hormone: increased sensitivity of the maturing follicle to FSH. Endocrinology 119: Bryant SM, Gale JA, Yanagihara DL, Campeau JD, DiZerega GS 1988 Angiogenic, mitogenic, and chemotactic activity of human follicular fluid. Am J Obstet Gynecol 158: Adashi EY, Resnick CE, Svoboda ME, van Wyk JJ 1984 Somatomedin C enhances induction of luteinizing hormone receptor by follicle-stimulating hormone in cultured rat granulosa cells. Endocrinology 116: Dodson WC, Schomberg DW 1987 The effect of transforming growth factor-/3 on follicle-stimulating hormone-induced differentiation of cultured rat granulosa cells. Endocrinology 120: Hillier SG, Trimbos Kemper TCM, Reichert Jr LE, van Hall EV 1983 Gonadotropic control of human granulosa cell function: in vitro studies on aspirates of preovulatory follicles In: Greenwald GS, Terranova PV (eds) Factors Regulating Ovarian Function. Raven Press, New York, pp Hillier SG, Wickings EJ, Afnan M, Margara RA, Winston RML 1984 Granulosa cell steroidogenesis before in vitro fertilization. Biol Reprod 31: Hillier SG, Harlow CR, Shaw HJ; Wickings EJ, Dixson AF, Hodges JK 1988 Cellular aspects of pre-ovulatory folliculogenesis in primate ovaries. Hum Reprod 3: Voss AK, Fortune JE 1993 Levels of messenger ribonucleic acid for cytochrome P450 17a-hydroxylase and P450 aromatase in preovulatory bovine follicles decrease after the luteinizing hormone surge. Endocrinology 132: Fitzpatrick SL, Richards JS 1991 Regulation of cytochrome P450 aromatase messenger ribonucleic acid and activity by steroids and gonadotropins in rat granulosa cells. Endocrinology 129: Ghersevich S, Nokelainen P, Poutanen M, Orava M, Autio- Harmainen H, Rajaniemi H, Vihko R 1994 Rat 17/3-hydroxysteroid dehydrogenase type 1: primary structure and regulation of enzyme expression in rat ovary by diethylstilbestrol and gonadotropins in vivo. Endocrinology 135: Chaffkin LM, Luciano AA, Peluso JJ 1992 Progesterone as an autocrine/paracrine regulator of human granulosa cell proliferation. J Clin Endocrinol Metab 75: Chaffkin LM, Luciano AA, Peluso JJ 1993 The role of progesterone in regulating human granulosa cell proliferation and differentiation in vitro. J Clin Endocrinol Metab 76: DiMattina M, Albertson B, Seyler DE, Loviaux DL, Falk RJ 1986 Effect of the antiprogestin RU486 on progesterone production by cultured human granulosa cells: inhibition of the ovarian 3/3-hydroxysteroid dehydrogenase. Contraception 34: Chandrasekher YA, Brenner RM, Molskness TA, Yu Q, Stouffer RL1991 Titrating luteinizing hormone surge requirements for ovulatory changes in primate follicles. II. Progesterone receptor expression in luteinizing granulosa cells. J Clin Endocrinol Metab 73: Greenberg LH, Stouffer RL, Brenner RM, Molskness TA, Hild- Petito SA, Yu Q 1990 Are human luteinizing granulosa cells a site of action for progesterone and relaxine? Fertil Steril 53: Richards JS 1994 Hormonal control of gene expression in the ovary. Endocr Rev 15: Matzuk MM, Finegold MJ, Su JGJ, Hsueh AJW, Bradley A 1992 a-inhibin is a tumour-suppressor gene with gonadal specificity in mice. Nature 360: Matzuk MM, Kumar TR, Vassalli A, Bickenbach JR, Roop DR, Jaenisch R, Bradley A 1995 Functional analysis of activins during mammalian development. Nature 374: Matzuk MM, Kumar TR, Bradley A1995 Different phenotypes for mice deficient in either activins or activin receptor type II. Nature 374: Matzuk MM, Lu N, Vogel H, Sellheyer K, Roop DR, Bradley A 1995 Multiple defects and perinatal death in mice deficient in follistatin. Nature 374: Chang PY, Benecke H, Le Marchand-Brustel Y, Lawitts J, Moller DE 1994 Expression of a dominant-negative mutant human insulin receptor in the muscle of transgenic mice. J Biol Chem 269: Ma YJ, Dissen GA, Merlino G, Coquelin A, Ojeda SR 1994 Overexpression of a human transforming growth factor-a (TGF-a) transgene reveals a dual antagonistic role of TGF-a in female sexual development. Endocrinology 135: Teerds KJ, Dorrington JH 1995 Immunolocalization of transforming growth factor a and luteinizing hormone receptor in healthy and atretic follicles of the adult rat ovary. Biol Reprod 52: